Decembe 2017 | Volume 8 | A icle 17091
Me hods
published: 08 Decembe 2017
doi: 10.3389/ immu.2017.01709
F on ie s in Immunology | www. on ie sin.o g
Edi ed by:
Ramesh Akkina,
Colo ado S a e Uni e si y,
Uni ed S a es
Re iewed by:
San hi Go an la,
Uni e si y o Neb aska Medical
Cen e , Uni ed S a es
Michael Scho sae ,
Icahn School o Medicine a
Moun Sinai, Uni ed S a es
*Co espondence:
Ha alampos Ha ziki ou
ha alampos.ha ziki ou@
helmhol z-hzi.de;
Michael Meye -He mann
mmh@ heo e ical-biology.de;
Rena a S ipecke
s ipecke. ena a@mh-hanno e .de
†Sha ed i s co-au ho ship.
Special y sec ion:
This a icle was submi ed o
Vaccines and Molecula
The apeu ics,
a sec ion o he jou nal
F on ie s in Immunology
Recei ed: 24Augus 2017
Accep ed: 20No embe 2017
Published: 08Decembe 2017
Ci a ion:
VolkV, ReppasAI, Robe PA,
SpineliLM, Sunda ase yBS,
TheobaldSJ, Schneide A,
Ge aschL, De esRo hC, KlössS,
KoehlU, Kaisenbe gC ,
Figuei edoC, Ha ziki ouH,
Meye -He mannM and S ipeckeR
(2017) Mul idimensional Analysis
In eg a ing Human T-Cell Signa u es
in Lympha ic Tissues wi h Sex o
Humanized Mice o P edic ion
o Responses a e Dend i ic
Cell Immuniza ion.
F on . Immunol. 8:1709.
doi: 10.3389/ immu.2017.01709
Mul idimensional Analysis In eg a ing
human -Cell signa u es in
Lympha ic issues wi h sex o
humanized Mice o P edic ion
o Responses a e dend i ic
Cell Immuniza ion
Vale y Volk1†, And eas I. Reppas2†, Philippe A. Robe 2, Loukia M. Spineli3,
Bala Sai Sunda ase y1, Sebas ian J. Theobald1, And eas Schneide 1, Lau a Ge asch1,
Candida De es Ro h1, S ephan Klöss4, Ul ike Koehl4, Cons an in on Kaisenbe g5,
Cons anca Figuei edo6, Ha alampos Ha ziki ou2*, Michael Meye -He mann2*
and Rena a S ipecke1*†
1 Depa men o Hema ology, Hemos asis, Oncology and S em Cell T ansplan a ion, Hanno e Medical School, Hanno e ,
Ge many, 2 Depa men o Sys ems Immunology, B aunschweig In eg a ed Cen e o Sys ems Biology, Helmhol z Cen e o
In ec ion Resea ch, B aunschweig, Ge many, 3 Ins i u e o Bios a is ics, Hanno e Medical School, Hanno e , Ge many,
4 Ins i u e o Cellula The apeu ics and GMP Co e Facili y IFB-Tx, Hanno e Medical School, Hanno e , Ge many, 5 Clinic o
Gynecology and Obs e ics, Hanno e Medical School, Hanno e , Ge many, 6 Depa men o T ans usion Medicine,
Hanno e Medical School, Hanno e , Ge many
Mice ansplan ed wi h human co d blood-de i ed hema opoie ic s em cells (HSCs)
became a powe ul expe imen al ool o s udying he he e ogenei y o human
immune econs i u ion and immune esponses in i o. Ye , analyses o human Tcell
ma u a ion in humanized models ha e been hampe ed by an o e all low immune eac-
i i y and lack o me hods o de ine p edic i e ma ke s o esponsi eness. Long-li ed
human len i i al induced dend i ic cells exp essing he cy omegalo i us pp65 p o ein
(iDCpp65) p omo ed he de elopmen o pp65-speci ic human CD8+ Tcell esponses
in NOD.Cg-Rag1 m1Mom-Il2 γ m1Wj humanized mice h ough he p esen a ion o immune-
dominan an igenic epi opes (signal 1), exp ession o co-s imula o y molecules (signal
2), and in lamma o y cy okines (signal 3). We exploi ed his alida ed sys em o e alua e
he e ec s o mouse sex in he dynamics o Tcell homing and ma u a ion s a us in
hymus, blood, bone ma ow, spleen, and lymph nodes. S a is ical analyses o cell
ela i e equencies and absolu e numbe s demons a ed highe CD8+ memo y Tcell
eac i i y in spleen and lymph nodes o immunized emale mice. In o de o unde s and
o which ex en he mul idimensional ela ion be ween o gan-speci ic ma ke s p edic ed
he immuniza ion s a us, he immunopheno ypic p o iles o indi idual mice we e used
o ain an a i icial neu al ne wo k designed o disc imina e immunized and non-
immunized mice. The highes accu acy o immune eac i i y p edic ion could be ob ained
om lymph node ma ke s o emale mice (77.3%). P incipal componen analyses u he
iden i ied clus e s o ma ke s bes sui ed o desc ibe he he e ogenei y o immuniza ion
esponses in i o. A co ela ion analysis o hese ma ke s e lec ed a issue-speci ic
2
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
impac o immuniza ion. This allowed o an o gan- esol ed cha ac e iza ion o he
immuniza ion s a us o indi idual mice based on he iden i ied se o ma ke s. This new
modali y o mul idimensional analyses can be used as a amewo k o de ining minimal
bu p edic i e signa u es o human immune esponses in mice and sugges s c i ical
ma ke s o cha ac e ize esponses o immuniza ion a e HSC ansplan a ion.
Keywo ds: hema opoie ic s em cell ansplan a ion, co d blood, dend i ic cell,  cell ma u a ion, lympha ic,
humanized mice, gende , a i icial neu al ne wo k
IN RodUC IoN
Humanized mice ansplan ed wi h human hema opoie ic s em
cells (HSCs) became a b oadly used expe imen al and p eclinical
pla o m o cha ac e ize he c i ical s eps o he econs i u ion
o he human immune sys em (1–3). In his con ex , humanized
mice a e cu en ly used o s udy human-speci ic in ec ions and
o es d ugs, accines, and cell he apies (2, 3). Eng a men o
human HSCs in he mouse bone ma ow (BM) and subsequen
ea ly Tcell de elopmen in hymus (Thy) could be con enien ly
s udied in sho - e m models las ing 10–16weeks (4). Ye , ull
ma u a ion o Tcells owa d memo y cells in HSC- ansplan ed
humanized mice was shown o be conside ably mo e he -
e ogeneous and challenging and equi ed pe iods o analyses
o 20weeks o longe (5, 6). Thus, his lymphopenia coincides
wi h he delayed Tcell immune econs i u ion in pa ien s a e
hema opoie ic s em cell ansplan a ion (HSCT) (5, 6). Mul iple
complemen a y app oaches we e ied o suppo he de elop-
men o human cells in immune-de icien mice such as, o
example, he adminis a ion o human cy okines (7) and he
gene a ion o new ansgenic mouse s ains exp essing human
cy okines (8) o human leukocy e an igens (HLA) molecules
(9, 10). Mo e complex and demanding s a egies explo ing
co- ansplan a ion wi h human e al hymus and li e issues
(bone ma ow, li e , hymus model) in o mice showed an o e all
imp o ed Tcell de elopmen and ma u a ion (11–14). No ably,
since Tcell esponses depend on he s eng h o he signals deli -
e ed by he an igen/HLA o he Tcell ecep o (TCR) (signal 1),
co-s imula ion (signal 2), and p o-in lamma o y cy okines
(signal 3), s udies demons a ing he p esence o human den-
d i ic cells (DCs) in humanized mice elucida ed hei ole in ac i-
a ion o he cogna e Tcells (15). Thus, as po en ial al e na i e
Abb e ia ions: aAPC, a i icial an igen-p esen ing cell; Ab, An ibody; ANN,
a i icial neu al ne wo k; ANOVA, analysis o a iance; B. D., below de ec ion;
BM, bone ma ow; BLT, bone ma ow, li e , hymus (mouse model); CB, co d
blood; CM, cen al memo y; CTL, cy o oxic lymphocy es; DN, double nega i e;
DP, double posi i e; DC, dend i ic cell; ELISPOT, enzyme-linked immuno spo
assay; EM, e ec o memo y; FBS, e al bo ine se um; HCMV, human cy omeg-
alo i us; HIV, human immunode iciency i us; HLA, human leukocy e an igen;
HSCs, hema opoie ic s em cells; HSCT, hema opoie ic s em cell ansplan a ion;
iDCpp65, induced dend i ic cells exp essing pp65; IDLV-G2a-pp65, in eg a ion-
de icien len i i al ec o co-exp essing GM-CSF/IFN-α and he HCMV-pp65;
IL, in e leukin; MCP-1, monocy e chemoa ac an p o ein 1; MLNs, mesen e ic
lymph nodes; N, naï e; NRG, NOD.Cg-Rag1 m1Mom-Il2 γ m1Wj; PB, pe iphe al
blood; PBS, phospha e-bu e ed saline; PCA, p incipal componen analysis; Pp65,
phosphop o ein 65; PLN, pe iphe al lymph nodes; SPL, spleen; TCRαβ, Tcell
ecep o αβ; TE, e minal e ec o ; Th1, T helpe ype 1; Thy, hymus; WT1, Wilms
umo 1 p o ein.
app oaches o imp o ing Tcell econs i u ion in humanized
mice and ul ima ely in humans, adop i e au ologous DCs, such
as hose explo ed clinically o cance immuno he apy (16)
and human immunode iciency i us (17), o in i o ac i a ed
DCs, as p e iously shown o be e ec i e in humanized mice
(18), could ep esen aluable op ions. Likewise, we ha e p e i-
ously desc ibed he p eclinical es ing o long-li ed gene ically
enginee ed induced DC (iDCs) in humanized mice. These cells
we e gene a ed a e a as o e nigh ansduc ion o monocy es
wi h len i i al ec o s encoding g anulocy e-mac ophage
colony s imula ing ac o (GM-CSF), in e e on-α (IFN-α), and
he human cy omegalo i us (HCMV) phosphop o ein (pp) 65
(19, 20). iDCs exp essing pp65 (iDCpp65) accines a e cu -
en ly in clinical de elopmen o p o ec ion o pos ansplan
pa ien s (21), since pp65 has been long known o be a majo
immune-dominan CD8+ cy o oxic Tlymphocy e a ge an igen
in heal hy se oposi i e adul s (22). Fu he mo e, non-exhaus ed,
long-li ed CD8+ e ec o memo y (EM) Tcells a e conside ed o
be c ucial o main ain li elong p o ec ion om HCMV eac i a-
ion in pos ansplan pa ien s (23).
We p e iously demons a ed ha mul iple adminis a ions
o iDCpp65 in o NOD.Cg-Rag1 m1Mom-Il2 γ m1Wj (NRG) mice
ansplan ed wi h human HSCs p omo ed a po en de elopmen
o CD8+ an igen-speci ic memo y esponses in sho (16weeks)
(20) and long (20–36 weeks) models (19, 24). We ha e also
demons a ed ha ano he impo an ac o o be conside ed
ega ding he analyses o human Tcells in mice humanized wi h
co d blood (CB)-HSCs is he gende o he ecipien mouse.
Fo he ini ial 10–15weeks a e HSCT, emales showed a mo e
obus Tcell de elopmen and ma u a ion, whe eas male’s Tcells
ma ched he emale’s Tcell ma u a ion s a us only 20weeks pos -
ansplan (25).
In his cu en wo k, we sough o e alua e whe he human-
ized emale and male mice would show di e en ial pa e ns
o Tcell esponses o iDCpp65. We cha ac e ized he CD4+/
CD8+ Tcells and hei subse s [naï e (N), EM, cen al memo y
(CM), and e minal e ec o (TE)] in di e en lympha ic issues
and con i med a dis inc beha io be ween emales and males,
suppo ed by s a is ical me hods. In o de o in eg a e he da a
ob ained om di e en issues and e alua e he immuniza ion
esponsi eness among hem, we adop ed a classi ica ion machine
lea ning algo i hm based on an a i icial neu al ne wo k (ANN).
A P incipal Componen Analysis (PCA) (26, 27) was u he
used o educe he c i ical in o ma ion equi ed o p edic
esponsi eness om he ANN (28). The ma ke s pinpoin ed by
he PCA e ealed ha he co ela ion s uc u e o o gan-speci ic
ma ke s is s ongly impac ed by immuniza ion and, he e o e,
3
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
iabili y (by ypan blue exclusion), len i i al copies pe cell
(by RT-q-PCR), and DC immunopheno ype (by low cy ome y)
we e pe o med as p e iously desc ibed (20, 21, 24). Fo immu-
niza ion, IDLV- ansduced cells we e hawed, washed wice wi h
PBS, and e-suspended in PBS a concen a ion 5.0×106cell/ml.
Cells we e kep on ice un il injec ion. A e CB-HSCT, mice we e
andomly dis ibu ed in o wo g oups, a non- ea ed con ol and
a g oup immunized wi h 5.0 × 105 iDCpp65 cells. Cells we e
adminis e ed subcu aneously in he le hind lank a weeks 6
and 10 and in igh hind lank a weeks 7 and 11 pos ansplan a-
ion (Figu e1A). Weekly weigh and gene al heal h moni o ing
we e pe o med un il he end o he expe imen a week 20
pos ansplan a ion.
s ep 3: Longi udinal Cha ac e iza ion
o human Cell de elopmen in PB
Samples o PB we e collec ed om all mice a weeks 10, 16, and
20 a e HSCT o e alua e he le el o eng a men o human
hema opoie ic cells and expansion o T and Bcells. Two ounds
o lysis we e pe o med o emo e e y h ocy es (0.83% ammo-
nium chlo ide/20mM HEPES, pH 7.2, o 5min a oom em-
pe a u e, ollowed by s abiliza ion wi h cold PBS and washing).
Cells we e labeled o low cy ome y analyses wi h he ollowing
an ibodies as p e iously desc ibed (24): paci ic blue an i-
CD45, Alexa700 (AF700) an i-CD19, allophycocyanin (APC)
an i-CD3, phycoe y h in-cyanine7 (PC7) an i-CD8 (BioLegend,
Fell, Ge many); allophycocyanin-H7 (APC-H7) an i-CD4,
Phycoe y h in (PE) an i-CD14 (BD Biosciences, San Jose, CA,
USA); and phycoe y h in-cyanine5 (PC5) an i-CD62L, luo es-
cein iso hiocyana e (FITC) an i-CD45RA (Beckman Coul e ,
K e eld, Ge many) o 15min a oom empe a u e, washed, and
analyzed by LSR II low cy ome e (BD Biosciences, Heidelbe g,
Ge many).
s ep 4: e alua ion o he human
Cell Responses agains pp65
in Immunized Mice
Th ee mice immunized wi h iDCpp65 and showing well-
de e loped lymph nodes we e used o ob aining human
memo y Tcells in high pu i y as p e iously desc ibed (20, 24).
C yop ese ed single cell suspensions gene a ed om lymph
nodes we e hawed, pooled, and e-suspended in X-VIVO
15 medium. Ac i a ion o Tcells was pe o med by MACS
magne ic beads conjuga ed wi h an i-CD2/CD3/CD28 mono-
clonal an ibodies (Mil enyi Bio ec, Ge many) in a bead- o-cell
a io o 1:2, in p esence o 25 IU/ml o human IL-2, 5 ng/
ml IL-7, and 5ng/ml IL-15 (Cellgenix, Ge many) o 48 h.
Cy okines we e e eshed e e y 2 days un il he end o he
homeos a ic expansion (day 9 pos -ac i a ion wi h magne ic
beads). Ac i a ed cells we e e-s imula ed a e cocul u e
ei he wi h au ologous iDC ( o APC-media ed homeos a ic
s imula ion bu lacking an igens) o wi h iDCpp65 o 7days
using a T/DC cell a io o 10:1. Fo in acellula IFN-γ de ec-
ion, expanded T cells we e i s seeded in iplica es wells
(3×105cells/well) o a 96-well ound-bo om pla e and hen,
o 16 h, he cells we e u he ac i a ed wi h 10 µg/ml o
ha hese ma ke s can be used as bioma ke s o e ie e he
in o ma ion o he immuniza ion s a us.
MA eRIALs ANd Me hods
s ep 1: Gene a ion o humanized Mice
ansplan ed wi h human CB-hsC
S udy p o ocols we e app o ed by he E hics Commi ee o he
Hanno e Medical School o acquisi ion and banking o human
HSCs ob ained om umbilical co d issues a e in o med
consen om dono s (mo he s a e m). The HSCs we e labeled
acco ding o a nume ical code ha could no be aced back
o he dono ’s pe sonal in o ma ion, hus keeping he dono ’s
anonymi y. All expe imen s in ol ing mice we e pe o med in
acco dance wi h he egula ions and guidelines o he animal
wel a e o he S a e o Lowe Saxony (Nds. Landesam ü
Ve b auche schu z und Lebensmi elsiche hei , Deze na 33/
Tie schu z). 5-week-old NRG mice we e o iginally ob ained
om The Jackson Labo a o y (JAX, Ba Ha bo , ME, USA) and
b ed in-house unde pa hogen- ee condi ions. P io o HSCT,
mice we e suble hally i adia ed (450 cGy) using a [137Cs]
column i adia o (Gammacell 3000 Elan; Bes The a onics,
O awa, ON, Canada). 4h a e i adia ion, 1.5–2.0×105 human
CD34+ hema opoie ic cells isola ed om emale dono umbili-
cal CB we e adminis a ed o each mouse ough he ail ein
as desc ibed (20, 24). We had p e iously shown ha immune
econs i u ion in emale mice ecipien s was as e han in
males (25) and we, he e o e, used emale dono s o a oid any
pu a i e immune esponses agains an igens exp essed in he Y
ch omosome o male ecipien s. S em cells om HLA*A02.01
posi i e (CB1, CB3) o nega i e (CB2) uni s we e used o gene -
a e humanized mice (CB1: n=11, CB2: n=10, CB3: n=9).
S a ing a week 10 pos ansplan a ion, he human immune
econs i u ion in mouse pe iphe al blood (PB) was assessed by
low cy ome y e alua ing he equency o human CD45+ cells.
s ep 2: Immuniza ion o Mice wi h idCs
exp essing he pp65 An igen
CD14+ monocy es we e isola ed a high pu i y ( om he same
CB uni s used as sou ce o CD34+ HSCs) by immune-magne ic
beads (Mil enyi Bio ec, Be gisch Gladbach, Ge many) and c yo-
p ese ed. CD14+ cells we e used o he gene a ion o iDCpp65
a e ansduc ion wi h a icis onic in eg ase-de ec i e len-
i i al ec o co-exp essing human cy okines GM-CSF/IFN-α
and he HCMV-pp65 p o ein as desc ibed (IDLV-G2a-pp65)
(20, 24). In sho , monocy es we e p e-condi ioned wi h ecom-
binan human GM-CSF and IL-4 (bo h 50 ng/ml; Cellgenix,
F eibu g, Ge many) o 8 h p io o len i i al gene ans e .
T ansduc ion o monocy es wi h IDLV-G2a-pp65 was pe o med
a a mul iplici y o in ec ion o 5 (2.5mg/ml p24 equi alen ) in
he p esence o 5µg/ml p o amine sul a e (Valean , Duesseldo ,
Ge many) o 16h. A e wa d, cells we e ha es ed by esuspen-
sion in phospha e-bu e ed saline (PBS), washed wice, and
c yop ese ed. Fo ansduc ion quali y assessmen , a sample
o ozen cells was hawed and main ained in he X-VIVO 15
medium (Lonza, Basel, Swi ze land) o 7days. Analyses o cell
FIGURe 1 | Gene a ion o induced dend i ic cells exp essing pp65 (iDCpp65) o immuniza ion o humanized mice. (A) Scheme o expe imen al design. Pu i ied
CD34+ hema opoie ic s em cells ob ained om h ee co d blood (CB) uni s and de oid o con amina ing Tcells we e used o ansplan a ion o h ee mice coho s.
Pu i ied CD14+ cells om he same CB we e ansduced wi h a icis onic in eg ase-de ec i e len i i al ec o co-exp essing huGM-CSF, huIFN-α, and HCMV-pp65.
C yop ese ed cells we e hawed, analyzed o iabili y, iden i y, and po ency cha ac e is ics in i o, and used o p ime/boos immuniza ions [a weeks 6, 7, 10, and
11 a e hema opoie ic s em cell ansplan a ion (HSCT), n=17]. Longi udinal analyses o pe iphe al blood we e pe o med on weeks 10, 16, and 20 a e HSCT.
Mice we e sac i iced a week 20 a e HSCT and bone ma ow, SPL, Thy, pe iphe al lymph node (PLN), and MLN we e isola ed and analyzed. Non-immunized
humanized mice (n=11) om he same co esponding CB uni s we e used as con ols. (B) Pe cen age o iable iDCpp65 cells a e c yop ese a ion and hawing
(whi e ba s) and 7days a e he in i o cul u e (black ba s) o each CB uni . (C) iDCpp65 gene a ed wi h CB 1 (whi e), 2 (g ay), and 3 (black ba s) we e main ained
o 7days in i o and he ex ac ed DNA was analyzed by RT-q-PCR o LV copy numbe pe cell. (d) Concen a ion o huIFN-α (g ay ba s) and huGM-CSF (black
ba s) de e mined o cell supe na an s collec ed a day 7 o in i o di e en ia ion o iDCpp65 gene a ed om CB dono 1 and 3 and measu ed by ELISA.
(e) Rep esen a i e do plo s o low cy ome y analyses o iDCpp65 (CB1) a haw and a day 7 o di e en ia ion in i o, showing high iabili y (7AAD nega i e
popula ion), down egula ion o CD14, up egula ion o CD45 and CD11c (used as ga es o u he analyses), and up egula ion o HLA-DR, CD80, CD86, and
pp65 upon iDCpp65 di e en ia ion.
4
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
CMV PepTi a o (pp65 o e lapping pep ide pool, Mil enyi
Bio ec, Be gisch Gladbach, Ge many) o 10µg/ml o Wilms
Tumo 1 (WT1) nega i e con ol o e lapping pep ide pool
(Mil enyi Bio ec, Ge many). A p o ein anspo inhibi o
cock ail (eBioscience, F ank u , Ge many) was added o he
cells 2h a e pep ide s imula ion. A he end o s imula ion,
he su ace s aining wi h APC an i-CD3, APC-H7 an i-CD4,
and PC7 an i-CD8 an ibodies (BioLegend, Ge many) was
5
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
pe o med. Subsequen ly, cells we e pe meabilized and ix-
a ed o in acellula s aining wi h PE an i-IFN-γ an ibodies
(eBioscience, San Diego, CA, USA). Samples we e acqui ed by
LSR II low cy ome e and da a we e analyzed using FlowJo
so wa e e sion 7.6.4 (T ee S a Inc., Ashland, OR, USA). As
a ge s o enzyme-linked immuno spo (ELISPOT) assays,
K562 cells exp essing HLA*A02.01 (also known as a i icial
an igen-p esen ing cells o “aAPCs”) and aAPCs exp essing
pp65 endogenously (aAPC/pp65) we e cul u ed in RPMI 1640
(Lonza, Swi ze land) con aining 10% e al bo ine se um. Fo
IFN-γ de ec ion by ELISPOT, Mul iSc een HTS pla es (Me k
Millipo e, Da ms ad , Ge many) we e coa ed wi h an i-IFN-γ
an ibodies (Mab ech, Nacka S and, Sweden) a 4°C o e nigh .
Then, 2.5×104 Tcells we e mixed wi h 7.5×104 aAPCs ei he
no ca ying an an igen, o wi h aAPCs pulsed wi h WT1
pep ides (aAPC+WT1), o wi h aAPCs pulsed wi h pp65 pep-
ides (aAPC+pp65) o exp essing pp65 endogenously (aAPC/
pp65). The Tcell/aAPC cocul u es we e incuba ed o e nigh ,
washed, and incuba ed wi h bio in-conjuga ed an i-human
IFN-γ monoclonal Ab, ollowed by incuba ion wi h alkaline
phospha ase-conjuga ed s ep a idin. Colo de elopmen was
pe o med using NBT/5-b omo-4-chlo o-3-indolyl phospha e
liquid subs a e, and he pla es we e analyzed in an ELISPOT
eade (AELVIS, Hanno e , Ge many). The mean ELISPOT
coun s o ac i a ion wi h iDC (4 eplica e cul u es) and
iDCpp65 (6 eplica e cul u es) we e ob ained.
s ep 5: Analyses o human Cy okines in
Mouse Plasma
A sac i ice, PB samples we e collec ed by hea punc u e and
cells we e subsequen ly sedimen ed by cen i uga ion. The supe -
na an con aining plasma was s o ed a −80°C un il he analysis.
A e hawing, plasma samples we e cen i uged a 2,000g o
10 min a oom empe a u e p io he analysis o emo e
emaining cell deb is. 25µl o plasma we e used o analysis o
each sample. The concen a ion in plasma o human GM-CSF,
IFN-γ, monocy e chemoa ac an p o ein (MCP-1), TNF-α,
IL-1β, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, and IL-12 (p70)
was analyzed by a 14-plex Luminex ki (Milliplex Millipo e, MA,
USA) acco ding o manu ac u e ’s p o ocol.
s ep 6: Cha ac e iza ion o he e minal
human Cell Responses in Lympha ic
issues
Spleen (SPL), pe iphe al lymph nodes (PLN), mesen e ic lymph
nodes (MLNs), BM, and Thy cells we e isola ed a e sac i ice and
homogenized. Fo mice ansplan ed wi h CB1-3, cell suspen-
sions gene a ed wi h PLN and MLN we e combined, whe eas o
CB2 and CB3, he e was an addi ional cell suspension ac ion
o MLN analyzed sepa a ely. Cell suspensions we e washed and
e-suspended in PBS o subsequen s aining wi h luo och ome-
conjuga ed monoclonal an ibodies as desc ibed abo e o PB
(24). Thymocy es we e s ained 15 min a oom empe a u e
wi h he ollowing an ibodies: paci ic blue an i-CD45, AF700
an i-CD4, APC an i-CD3, PC7 an i-CD8, FITC an i-TCR αβ
(BioLegend, Ge many). A e washing, cells we e e-suspended
in PBS con aining 1% human se um and analyzed by an LSR II
low cy ome e .
s ep 7: es ablishing he da abase and
s a is ical Analyses
All a iables consis ing o cell pheno ypes de e mined as ela i e
equency (PB and lympha ic issues) and coun s (lympha ic
issues) and cy okines concen a ions (plasma) we e o ganized
in a Pi o Table using Excel so wa e 2010 (Mic oso , Redmond,
WA, USA). Each sample was coded acco ding o he CB uni
(1, 2, o 3) cell ype, issue, ime-poin o analyses, in e en-
ion g oup (con ol o iDCpp65), and mouse gende ( emale
“F” o male “M”). Be a eg ession analysis was employed o
model he associa ion be ween he in e en ion g oups and
he cell pheno ypes measu ed as “ ela i e equency,” which is
exp essed as odds a io (i.e., he odds o e en be ween iDCpp65
and con ol), whe eas nega i e binomial eg ession was used
o model he associa ion be ween he in e en ion g oups
and he coun cell pheno ypes exp essed as a e a io (i.e., he
incidence a e be ween iDCpp65 and con ol). Bo h models
whe e implemen ed wi h and wi hou s a i ica ion by gende .
Pa ame e es ima ion was pe o med by leas squa e means. The
p- alues calcula ed a signi icance le els 0.05 and 0.01 using
he wo-sided z- es s a is ic we e conside ed signi ican . All
analyses we e implemen ed using he SAS 9.3 so wa e (SAS,
Ca y, NC, USA). PROC GLIMMIX was used o Be a eg ession
analysis, oge he wi h PROC GENMOD o he nega i e bino-
mial eg ession analysis. A wo-way analysis o a iance was
pe o med o analyze he esul s o he IFN-γ ELISPOT assay
using G aphPad P ism e sion 5 so wa e (G aphPad So wa e,
Inc., La Jolla, CA, USA).
s ep 8: ANN Classi ica ion App oach
The compu a ional analysis was ca ied ou in Ma Lab e sion
7.11.0, 2010 (Ma hWo ks, Inc., Na ick, MA, USA) using he
Neu al Ne wo k Pa e n ecogni ion applica ion. The inpu s
(15 ma ke s ep esen ing cell pheno ypes o PB, SPL,
PLN+MLN, o MLN, BM and 8 ma ke s o Thy) and hei
co esponding ou pu s [immuniza ion s a us o each mice,
“yes” (1) o “no” (0)] we e used o aining he ANN. To ind
he app op ia e numbe o neu ons in he hidden laye , we
pe o med a k- old c oss alida ion (29) (wi h k being equal
o 3, 4, 5) esul ing in 12 neu ons (8 o Thy). The ou pu laye
consis ed o wo neu ons. The classi ica ion accu acy (pe cen -
age o co ec classi ica ions) was es ima ed by a e aging o e
2000 ANN. Fo each o hose ANN, he inpu da ase was
andomly di ided in o aining (70% samples/ issue), alida-
ion (15% samples) and es ing (15% samples). The Le enbe g–
Ma qua d back-p opaga ion ( ainlm) algo i hm was used o
he aining s ep, as i is bes sui ed o small ne wo ks (30).
Fo he hidden laye , we used sigmoid ans e unc ions. The
da a we e i s s anda dized (i.e., each ma ke was no malized
o 0 mean alue and uni a iance) in o de all he ma ke s o
be a he same scale.
We also pe o med he same classi ica ion analysis o immu-
nized (1) and non-immunized mice (0), bu by di iding ou

6
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
samples in o male and emale mice da a se s. In his case, we
used he same pa ame e se ing o ANN as abo e and es ed
he classi ica ion accu acy o each subpopula ion. In addi ion
o he classi ica ion accu acy (de ined as he ac ion o co ec ly
classi ied samples), we measu ed he sensi i i y and speci ici y
o he classi ica ion, applied inside each g oup. The sensi i i y,
de ined as
sensi
i i y=
numbe o ue posi i es
numbe o ue posi i ees + numbe o alse nega i es ,
ep esen s he p obabili y o a sample classi ied as immunized
(1) o belong o he immunized g oup.
The speci ici y p o ides he p obabili y o a sample classi ied
as non-immunized (0) o belong o he non-immunized g oup.
Speci ici y is de ined as
sp
eci ici y=
numbe o ue nega i
numbe o ue nega i es
ees + numbe o alse posi i es .
The 15 ma ke s co espond o he equencies o human lym-
phocy e lineages analyzed pe issue (PB, BM, SPL, MLN+PLN,
MLN) o each mouse included cells de e mined as equencies
o CD45, CD19/CD45, CD3/CD45, CD14/CD45 (in BM CD34/
CD45 ins ead o CD14/CD45), CD4/CD45, CD8/CD45 and a
equency o o he non-de e mined CD45+ cells, equencies o
CD4 sub ypes (N, CM, EM, and TE), and equencies o CD8
sub ypes (N, CM, EM, and TE). Eigh ma ke s we e used o
Thy: equencies o CD45, CD3/CD45, CD4/CD45, CD8/CD45,
CD4−/CD8− (double nega i e, DN), CD4+/CD8+ (double posi-
i e, DP), TCRαβ+/CD3, and CD4 o CD8 a io alues.
s ep 9: PCA
P incipal componen analysis (31) was used o ecognize
clus e s o in e ela ed ma ke s o con ol and iDCpp65 mice
in he di e en issues. PCA cons uc s speci ic di ec ions,
which a e called p incipal componen s, along which he da a
a e mos dispe sed and hus bes dis inguishable. In his way, a
da a se can be ep esen ed by he p incipal componen s, which
inco po a e a speci ic amoun o he a iance (o dispe sion).
Since he componen s a e unco ela ed o each o he , he ma k-
e s, which a e s ongly co ela ed wi h a componen , compose
a clus e o ma ke s, which a y oge he . He e, we c ea ed
hese clus e s by selec ing he pheno ypic ma ke s ha we e
s ongly co ela ed o an i-co ela ed (mo e han 80%) wi h
any componen o each g oup. The i s ou componen s we e
conside ed o analysis since hey we e able o inco po a e mo e
han 80% o he o al a iance in bo h con ol and iDCpp65
mice in all issues. These componen s we e he basis o a iance-
dis ibu ion compa isons and co ela ion hea -maps o bo h
mouse g oups. The ma ke s ha we e s ongly co ela ed o
an i-co ela ed wi h he i s go e ning componen we e used
o co ela ion compa isons be ween he con ol and iDCpp65
mice. The da a o con ol and iDCpp65 mice om BM and
SPL we e ep esen ed by 30 a iables (pe cen ages and coun s
o each o he 15 ma ke s); om combined PLN, PB, and MLN
by 15 a iables (only pe cen ages); om Thy by 14 a iables
(pe cen ages and coun s). As in he aining o ANN, he da a
we e i s s anda dized (i.e., each ma ke was no malized o 0
mean alue and uni a iance) in o de eno malize all ma ke s
o he same scale.
ResUL s
C yop ese ed idCpp65 Remained
Viabili y and Cha ac e is ics a e hawing
We ha e shown be o e ha human T cell esponses agains
HCMV-pp65 we e consis en ly s imula ed in humanized NRG
mice immunized wi h iDCpp65 (19, 20, 24). He e, iDCpp65 we e
gene a ed and c yop ese ed immedia ely a e ansduc ion, and
subsequen ly used o p ime-boos immuniza ions o emale and
male mice (Figu e1A). E ec s o immuniza ions in lympha ic
issues we e analyzed 20 weeks a e HSCT co esponding o
9weeks a e he las immuniza ion. The cell accine, iDCpp65,
showed high iabili y di ec ly a e hawing (66–88%) and in i o
cul u e o 7days (42–63%) ela i e o s a ing numbe o cells
(Figu e1B). E icien ansduc ion wi h IDLV and pe sis ency o
episomal i al copies we e con i med o cells main ained in cul-
u e o 7days and showing in a e age i e len i i al copies/cell
(Figu e1C). T ansgenic cy okines ha accumula ed on he cell
supe na an o iDCpp65 de i ed om CB1 and CB3 o 7days
we e de ec ed in he ange o 500–1,000pg/ml o IFN-α and
50–100pg/ml o GM-CSF (Figu e1D). The iable cells showed
a ypical DC immuno pheno ype wi h co-exp ession o HLA-DR
(61.60–91.85%), CD86 (93.20–98.97%), and CD80 (29.1–97.7%)
su ace ma ke s (Figu e 1E, ep esen a i e da a o a ba ch o
iDCpp65, Figu e S1A in Supplemen a y Ma e ial). In acellula
immunos aining o de ec ion o pp65-posi i e iDCpp65 cells by
low cy ome y assay showed a iable esul s among di e en CB
dono s (CB1: 49.20%, CB2: 2.49%, CB3: 16.40% when calcula ed
o o al iable cells in suspension 7days a e in i o cul u e,
Figu e S1B in Supplemen a y Ma e ial).
idCpp65 Immuniza ions A ec ed
Lymphocy es Coun s in PB, Plasma
Cy okines P o iles, and Lymph Node
de elopmen
Th ee independen coho s o NRG mice a e CD34+ HSCT we e
gene a ed. 5.0×105 hawed and iable au ologous iDCpp65 we e
injec ed a weeks 6, 7, 10, and 11 a e HSCT. Body-weigh and
gene al heal h condi ions we e moni o ed weekly. Al hough
emales om bo h con ol and iDCpp65-immunized coho s
we e ligh e han males, mice o bo h gende s gained weigh
no mally o he 20weeks a e HSCT (Figu e2A) and showed
no signs o g a - e sus-hos disease o o gan pa hologies a
sac i ice (da a no shown). The equencies o human CD45+
cells de ec ed in PB a weeks 10, 16, and 20 pos -HSCT we e
consis en ly highe in emales, pa icula ly o emales immu-
nized wi h iDCpp65 (Figu e2B). The expansion o human CD8+
and CD4+ Tcell de ec able in PB was supe io in he iDCpp65
coho compa ed o con ols, especially in he male g oup
(Figu e2B; Table S1 in Supplemen a y Ma e ial). Analyses o
human cells in PB 20weeks a e HSCT by assessmen o he
FIGURe 2 | Longi udinal analyses o emale and male humanized mice. (A) Weigh moni o ing o con ol (n=11, uppe panel) and immunized mice (n=17,
lowe panel). A ows indica e he weeks o induced dend i ic cells exp essing pp65 (iDCpp65) immuniza ions a e hema opoie ic s em cell ansplan a ion. Bi-weekly
weigh (g) de e mined o emales (F) indica ed in g ay and o males (M) in black. The numbe o F and M mice pe g oup is indica ed. (B) Mean ela i e equencies
o human CD45+, CD8+ in CD45+, and CD4+ in CD45+ cells de e mined in blood by low cy ome y o con ol (dashed line) and iDCpp65-immunized (solid line)
F (uppe panel) and M (lowe panel) a weeks 10, 16, and 20 pos ansplan a ion. E o ba s ep esen SEs. *p<0.05 a e indica ed on he g aph. (C) Rela i e
equencies o human cell ypes measu ed by low cy ome y in PBL o emale (F) and male (M) mice a week 20 pos ansplan a ion. Mean ela i e equencies o
CD19+ (whi e), CD4+ (ligh g ay), CD8+ (da k g ay), and o he CD45+ cells (black) we e measu ed in con ol and iDCpp65-immunized mice, *p<0.05 a e indica ed
on he g aph. (d) Pheno ypes o dis inc CD8+ and (e) CD4+ Tcells we e de e mined as naï e (N, whi e, CD45RA+/CD62L+), cen al memo y (CM, ligh g ay,
CD45RA−/CD62L+), e ec o memo y (EM, da k g ay, CD45RA−/CD62L−) and e minal e ec o (TE, black, CD45RA+/CD62L−). Mean ela i e equencies a e shown
o con ol and iDCpp65-immunized mice. (F) Concen a ion o human cy okines measu ed in plasma o F (uppe panel) and M (lowe panel) mice 20 weeks
pos ansplan a ion. Mean concen a ions o cy okines we e de e mined o iDCpp65-immunized (g ay) and con ol (whi e) g oups. Ba s and ci cles e lec he SE o
he es ima ed mean concen a ions and he obse ed concen a ions o indi idual samples, espec i ely. Concen a ion alues below he de ec ion limi o he assay
a e indica ed (b.d.), *p<0.05, **p<0.01 indica ed on he g aph. (G) The a e age equency o lymph nodes ound in F ( op panel) and in M (bo om panel) mice o
iDCpp65-immunized and con ol g oups. Values o inguinal (uppe -), axilla y (middle-), and iliac (lowe panel) lymph nodes ound 20weeks a e hema opoie ic s em
cell ansplan a ion. Numbe o samples: emale, n=5/9; male, n=6/8, con ol/immunized, espec i ely.
7
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
FIGURe 3 | Func ional memo y T esponses agains pp65 a e induced
dend i ic cells exp essing pp65 (iDCpp65) immuniza ion. (A) Tcells isola ed
om lymph nodes o immunized mice (n=3) we e e-s imula ed in i o wi h
ei he induced DC (iDC) o iDCpp65. Absolu e Tcell numbe s be o e (whi e
ba s) and 7days a e (black ba s) s imula ion a e shown. (B) Rela i e old
inc ease o cells popula ions be o e and 7days a e cocul u e wi h iDC (g ay
ba s) and iDCpp65 g oup (black ba ). (C) CD8+ and (d) CD4+ Tlymphocy es
expanded a e cocul u e wi h iDC (n=3) o iDCpp65 (n=3) we e le
uns imula ed (whi e ba s) o e-s imula ed wi h a Wilms Tumo 1 (WT1)
pep ide pool (g ay ba s) o pp65-pep ide pool (black ba s). F equency o cells
p oducing IFN-γ is shown. (e) Tcells expanded wi h iDC ( ou cul u es pe
g oup) o iDCpp65 (six cul u es pe g oup) we e cocul u ed wi h an a i icial
an igen-p esen ing cell (aAPC) on an ELISPOT pla e. The aAPCs we e e he
no loaded (whi e ba s), loaded wi h WT1 pep ides (ligh g ay ba s), loaded
wi h pp65-pep ides (da k g ay ba s), o ansduced o endogenous pp65
exp ession and loading (black ba s). Mean absolu e numbe o IFN-γ-spo s
and *p<0.05 (analysis o a iance), †p<0.05 (F- es ) a e indica ed.
8
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
mean ela i e equencies showed ha CD8+ Tcells we e highe
in he immunized coho (p=0.03) (Table S1 in Supplemen a y
Ma e ial), wi h high signi icance o he male g oup (p=0.01
CD8+; p=0.09 CD4+) (Figu e2C; Table S1 in Supplemen a y
Ma e ial). This was associa ed wi h a highe accumula ion o EM
and TE CD8+ cells in he g oup o iDCpp65-immunized males
whe eas he g oup o emales showed highe accumula ion o
CM CD8+ cells in immunized e sus con ol g oup (Figu e2D;
Table S1 in Supplemen a y Ma e ial). The ela i e equencies
o CD4+ Tcell sub ypes we e only sligh ly al e ed, showing an
inc ease in he ela i e equencies o CM cells o males and
EM cells o emales (Figu e 2E; Table S1 in Supplemen a y
Ma e ial). A luo escen -based bead assay was used o measu e
he concen a ion o 12 human cy okines in mouse plasma
(GM-CSF, MCP-1, IFN-γ, TNF-α, IL-1β, IL-2, IL-4, IL-5, IL-6,
IL-8, IL-10, and IL-12p70). IL-1β, IL-2, IL-4, and IL-12p70 we e
below he de ec ion limi . Fo he g oup o non-immunized mice,
he baseline concen a ions o human cy okines in he plasma
we e consis en ly highe o males. Rema kably, IL-5, MCP-1,
and IL-8 we e only de ec able in non-immunized males. Upon
immuniza ions, bo h gende s showed inc eased IFN-γ ( a e a io
2.74, p= 0.15) and GM-CSF concen a ions ( a e a io 2.96,
p=0.092) (Table S2 in Supplemen a y Ma e ial). The inc ease o
IFN-γ concen a ion a e immuniza ion was pa icula ly high
o immunized emales (p=0.003) (Figu e2F). De ec ion o
IL-5, MCP-1, IL-6, and IL-8, in plasma o emales was only pos-
sible a e iDCpp65 immuniza ion. The e o e, o e all, iDCpp65
immuniza ion ha nessed he ma u a ion o human T cells in
PB, which was in gene al associa ed wi h an inc ease o human
cy okines in plasma (in pa icula o emales). The unde sized
and incomple ely de eloped lymph nodes in humanized mice
a e di icul o be de ec ed. They can be, none heless, de ec ed as
qui e small “ a y” s uc u es in he expec ed ana omical egions
(inguinal, axilla y, iliac). Upon immuniza ion wi h iDCpp65,
hese d aining LNs become mac oscopically mo e no iceable.
Al hough hese egene a ed LNs a e no ully “no mal” in ela-
ionship o lymph nodes ound in immune compe en mice, hey
con ain a high densi y o human Tcells (19). In his cu en s udy,
we con i med a highe equency o de eloped d aining lymph
nodes (nea he immuniza ion si es) in he immunized coho s,
which was mo e e iden o males (Figu e 2G). Rema kably,
his obse a ion was in e ed o he axilla y lymph nodes,
which we e mo e p ominen in emales. On he o he hand,
he a e age equency o de ec able iliac nodes was lowe in he
immunized coho (Figu e2G). Small MLNs de eloped in mos
mice, ega dless o gende o immuniza ions (da a no shown),
indica ing he possibili y ha he de elopmen o MLNs may be
induced di e en ly.
idCpp65-Immunized Mice demons a ed
Func ional pp65-speci ic Memo y
Cell Responses
Lymph nodes o immunized mice we e a aluable compa -
men o he de ec ion o high equencies o human T cells.
Lymphocy es eco e ed om lymph nodes o h ee immu-
nized mice we e pooled and main ained in i o o 48h o
homeos a ic ac i a ion by beads and cy okines. Lymphocy es
om non-immunized mice we e no used since, om p e ious
expe ience, we we e no able o expand hem success ully in i o
(19, 20). Mic ocul u es o cell suspensions we e incuba ed wi h
au ologous “emp y” iDCs o wi h iDCpp65 a 10:1 T o DC a io
o 1week o p omo e u he expansion o Tcells. The expan-
sion was mo e p onounced o Tcells cul u ed in he p esence
o iDCpp65 s imula ion han “emp y” iDCs ( old expansion
ela i e he popula ion be o e s imula ion): 10.9 (iDCpp65) and
9.1 (iDCs) (Figu es3A,B). Ac i a ed CD8+ and CD4+ Tcells
we e analyzed by low cy ome y o de ec ion o in acellula
IFN-γ. Tcells expanded in he p esence o emp y iDC and e-
s imula ed wi h WT1 o pp65-pep ide pools showed simila
baseline equencies o CD8+ IFN-γ+ and CD4+ IFN-γ+ Tcells.
In con as , Tcells expanded in he p esence o iDCpp65 and
hen e-s imula ed wi h he pp65-pep ide pool showed a iable
bu in a e age much highe ela i e equencies o CD8+ IFN-γ+
FIGURe 4 | Rela i e and absolu e quan i ica ion o human hema opoie ic lineages in lympha ic issues. Mean ela i e equency o human CD19+ (whi e), CD4+
(ligh g ay), CD8+ (da k g ay), and o he CD45+ cells (black) de ec ed by low cy ome y analyses o (A) bone ma ow, (B) SPL, (C) combined pe iphe al lymph nodes
and MLN, and (d) MLN. F equencies o emale (F) and male (M) mice a e shown sepa a ely. (e) Absolu e cell coun s o bone ma ow and (F) spleen we e ob ained.
Plo s ep esen a e age numbe o cells de e mined o con ol (g ay ba s) and induced dend i ic cells exp essing pp65 (iDCpp65)-immunized (black ba s) spli
be ween emale and male mice. The numbe o mice analyzed pe g oup (n) and *p<0.05 is indica ed on he g aph. Fo (A,B,e,F) F/con ol n=5, F/iDCpp65
n=9, M/con ol n=6, M/iDCpp65 n=8. Fo (C) F/con ol n=3, F/iDCpp65 n=6, M/con ol n=4, M/iDCpp65 n=6. Fo (d) F/con ol n=4, F/iDCpp65 n=5,
M/con ol n=6, M/iDCpp65 n=6.
9
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
and CD4+ IFN-γ+ Tcells (Figu es3C,D). The mean an i-pp65
esponse o CD8+ was 6.84 imes highe in he iDCpp65
han in iDC e-s imula ion g oup and o CD4+ Tcells in 6.27
imes, espec i ely. F- es compa ing a iances be ween iDC
and iDCpp65 showed a s ong e idence o a iance di e ence
be ween g oups in case o e-s imula ion wi h pp65-pep ide pool
(p=0.025, CD8+; p=0.072, CD4+) bu no when g oups we e
e-s imula ed wi h WT1 o no s imula ed a all (p>0.05 o all
cases) (Figu es3C,D).
As a complemen a y app oach, he abili y o Tcells o ec-
ognize and be ac i a ed by pp65 epi opes p esen ed by an aAPC
posi i e o HLA-A*02.01 was es ed by an IFN-γ-ELISPOT
assay as p e iously desc ibed (32). Tcells expanded a e cocul-
u e wi h iDCs o iDCpp65 we e exposed o e nigh o di e -
en ypes o aAPCs, and he numbe s o eac i e Tcells we e
quan i ied. Tcells expanded wi h iDCpp65 and cocul u ed wi h
ei he aAPC loaded wi h pp65-pep ides o ansduced o pp65
exp ession showed on a e age signi ican ly highe equencies
o ac i a ed T cells han when cocul u ed wi h aAPC loaded
wi h con ol WT1 pep ides (p=0.016, p=0.026, espec i ely).
No signi ican ampli ica ion o Tcell ac i a ion was obse ed
when Tcells we e expanded in he p esence o “emp y” iDCs
(p>0.05 o bo h cases) (Figu e3E). These da a con i med
ha immuniza ions wi h iDCpp65 p omo ed a speci ic immune
compe ence agains pp65 in humanized mice which was medi-
a ed by human Tcells.
he e ogeneous Pa e ns o human
Lymphocy es in Lympha ic issues
a e idCpp65 Immuniza ion
Isola ed issues [BM, SPL, lymph nodes (PLN+MLN and MLN)]
we e p ocessed o low cy ome y analyses and quan i ica ion
o human lymphocy e equencies and absolu e numbe s. Fo
BM, CD3+ T cells ep esen ed only a mino i y o huCD45+
cells, whe eas Bcells (CD19+) and o he CD45+ cells p e ailed.
In e ms o ela i e equencies, a end o inc eased equencies
o CD19+ cells was obse ed upon iDCpp65 immuniza ion,
16
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
The same analysis in MLN e ealed impo an di e ences
in he a iance dis ibu ion in con ol and iDCpp65 mice
(Figu e8E) and conside able co ela ion changes in he hea -
map analysis (Figu e9E). In e se co ela ed signa u es could
be seen be ween CD4% and CD8CM% (posi i e co ela ion
in con ol mice, equal o 0.5969 and nega i e co ela ion in
iDCpp65 ones, equal o −0.618), CD4% and CD8CM% (posi-
i e co ela ion in con ol mice, equal o 0.8102, and a weakly
nega i e co ela ion in iDCpp65 ones, equal o −0.384). The
same in e sely co ela ed signa u e could be seen be ween
CD4% and CD8N% (nega i e co ela ion in con ol mice, equal
o −0.6346 and posi i e co ela ion in iDCpp65 ones, equal o
0.7513).
Finally, in PB, he same analysis e ealed a s iking di e -
ence in he a iance dis ibu ion (Figu e8F) and conside able
di e ences in he hea -map analysis (Figu e 9F). In e sely
co ela ed signa u es could be seen be ween CD4CM% and
CD4EM% (posi i e co ela ion in con ol mice, equal o 0.6414
and nega i e co ela ion in iDCpp65 ones, equal o −0.598) as
well as be ween CD4EM% and CD8CM% (posi i e co ela ion
in con ol mice, equal o 0.6567 and weakly nega i e co ela-
ion in iDCpp65 ones, equal o −0.3047). The posi i e s ong
co ela ion be ween CD4EM% and CD8EM% in con ol mice
(equal o 0.94) became neu al in iDCpp65 (−0.09983) and he
nega i e s ong co ela ion be ween CD19% and CD4EM% in
con ol mice (equal o −0.8088) becomes neu al in iDCpp65
(−0.06359).
The abo e esul s showed how we could exploi he in insic
complexi y and he e ogenei y o he inpu ma ke s o gain
addi ional knowledge on he cha ac e is ic o an indi idual
immunized mouse. Mo e speci ically, by checking he co ela ion
pa e ns among speci ically selec ed ma ke s in an indi idual
mouse, we could conclude abou hei immuniza ion s a us.
Toge he wi h he classi ica ion pe o mance o ANN in he di -
e en issues, hese esul s p edic ed mo e p onounced e ec s o
immuniza ion in SPL and PLN combined wi h MLN compa ed
wi h o he issues.
dIsCUssIoN
The use o immune de icien mice humanized wi h human HSCs
o s udy and cha ac e ize he ma u a ion o human Tcells a e
immuniza ions in di e en lympha ic issues gene a e la ge
da a se s and highly complex esul s. Long- e m s udies (20 o
mo e weeks a e ansplan a ion) and la ge mouse coho s
(15 mice o mo e) ha e been commonly used (6). The ini ial HSC
eng a men in BM, ea ly Tcell de elopmen in hymus and he
eg ess o naï e Tcells o he pe iphe y ecapi ula e he gene al
pa e ns ound in immune compe en mice and in humans
(5, 33, 34) in he i s 10–15 weeks a e ansplan a ion.
Howe e , analyses o he T cell ma u a ion in seconda y
lymphoid o gans ha e shown o be mo e he e ogeneous and
p edic i e o he quali y o he immune econs i u ion and
de elopmen o ma u e T cells. This e lec s he “pe sonal-
ized” condi ion o di e en CB dono s wi h he e ogeneous
gene ic backg ounds, which is ampli ied in a xenog a sys em.
In addi ion, al hough he eng a men o human HSCs (35)
and he highe hymic ou pu in humanized emale mice (25)
had been p e iously epo ed, he ele ance o he mouse sex
in de e mining he impac o immuniza ions and he p edic -
abili y o human Tcell esponses had no been p esen ed. All
hese ac o s we e aken in accoun when explo ing humanized
mice o es ing a new accine ype.
In he p esen wo k, we sough o e alua e he mul idimen-
sional spa ial e ec s o a po en cellula accine agains HCMV
ma ched o he HSC dono and p o iding he h ee main ele an
signals o bo h an igenic and homeos a ic ac i a ion o Tcells:
an immune-dominan an igen p esen ed ia HLA class I and II,
co-s imula o y ligands, and in lamma o y cy okines. Thus,
iDCpp65 which a e iable o 2–3weeks in i o and e ec i ely
mig a e o lymph node s uc u es (19) we e used o accele a e
and po en ly boos he human Tcell de elopmen and unc ional
esponses in humanized mice. As an icipa ed, imp o ed human
Tcell de elopmen and ma u a ion we e longi udinally obse ed
in PB and e minally in se e al lympha ic issues in iDCpp65-
immunized mice a 20-weeks a e HSCT. These esul s comple-
men ed p e ious indings ob ained in sho e (16weeks) and in
longe (up o 36weeks) iDCpp65 immuniza ion models (20, 24).
Ad anced s a is ical analyses showed ha iDCpp65 immuniza-
ions p omo ed a ypical memo y Tcell signa u e (abo e all o
CD8+ Tcells) which was mos p ominen o Tcells homing
lymph nodes and spleen. As mos o he s udies using human-
ized mice ha e ocused on analyses o human cells in blood and
spleen (5, 8, 18) i is impo an o emphasize ha , as seen om
he ANN and PCA analyses, he quan i y and quali y o human
Tcell eac i i y in pe iphe al and MLNs (e en i hey a e small
and di icul o be sampled) ha e o be aken in accoun , as
lymph nodes ep esen he p ime issue o in e ac ions be ween
an igen-p esen ing cells wi h naï e CD8+ and CD4+ T cells.
Fu he , he le els o human IFN-γ in plasma inc eased upon
immuniza ion.
In gene al, bo h cellula and cy okine immune e ec s we e
mo e accen ua ed o emale mice. This con i med and expanded
ou p e iously epo ed obse a ion ha humanized emale
mice ha e highe ou pu o naï e Tcells han humanized male
mice un il 12weeks a e HSCT (25). A ound 16weeks a e
HSCT, male mice showed highe de elopmen o ma u e Tcells
and by 20weeks a e HSCT, he equencies o human CD45+
cells, naï e and memo y Tcells in PB equalized be ween he
sexes. No a e al. showed ha be ween 10 and 12weeks a e
HSCT, emales ansplan ed wi h limi ing amoun s o HSCs
ob ained om se e al CB uni s gene ally exhibi ed a highe
equency o huCD45+ cells han male mice (35). The e o e,
CB-HSCT in humanized mice could po en ially mi o he
e ec o sex s e oids on human immune econs i u ion since
empo a ily blocking sex s e oids be o e HSCT in pa ien s,
inc eased hymus unc ion and enhanced he a e o T-cell
egene a ion (36). Responses o a ious ypes o accina ion a e
o en highe among women [ o a e iew see Re . (37)], who a e
able o moun s onge humo al esponses han men. A possible
explana ion o his phenomenon is based again on majo sex
s e oid ho mones such as he ypical “ emale” ho mone es adiol
ha enhances he adap i e and inna e immune sys ems, and he
“male” ho mone es os e one conside ed immune supp essi e.

17
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
No ewo hy, women display highe T helpe ype 2 (Th2)
esponses, whe eas males a o Th1 esponses (37). Al hough
sex-speci ic esponses o dis inc accines a e no usually con-
side ed and ha e been epo ed in a ew clinical ials, his is an
impo an ac o also o be conside ed in p eclinical esea ch,
when es ing new accine ypes, including when humanized
mice a e used as a po ency model. In he cu en model, we
ob ained no only highe esponses, bu explo ing he ANN, also
a be e p edic ion o esponse. Thus, as a logical app oach o
educe he numbe s o humanized mice when es ing a accine
is ini ially a o ing he use o emale mice. In addi ion, om
now on, s udies on humanized mice should conside male and
emale esponses as dis inc esponses, and should be analyzed
sepa a ely and compa ed.
We also showed ha he s a is ical me hods can be com-
plemen ed wi h an ANN algo i hm in o de o pin down he
complexi y o a mul idimensional da a se s including usual
immune ma ke s such as equencies o he human cell pheno-
ypes among lympha ic issues and conside ing mouse sexes.
As gene ally p oposed o ANNs (38), we we e able o demon-
s a e he e ha he ANN based on he humanized mouse da a
could “lea n” o ecognize he immune p ope ies o immunized
e sus con ol mice. Fo s udies in humans, ANNs we e buil
wi h independen immunologic a iables such as cell p oli -
e a ion, pheno ypic ma ke s, and cy okine exp ession in he
con ex o p os a e cance and in HSCT pa ien s (39–41). To ou
knowledge, he applica ion o ANNs o humanized mouse mod-
els o p edic ing he accu acy o immune esponse o de ining
signa u es o Tcell esponses was no p e iously pe o med. The
iden i ica ion o lymph node and spleen as he mos p edic i e
o gans o he immune s a e o con ol e sus immunized mice
migh be u he imp o ed by eleasing he assump ion o s a is-
ically independen issues. Along wi h local immune popula ion
dynamics, i would be o in e es o in es iga e he immune cell
a icking dynamics be ween di e en issues (42, 43). In his
way, we could elax he assump ion o issue independence.
Howe e , he immune cell a icking is an open challenge ye
o be sol ed in u u e esea ch.
Al oge he , he cu en app oach, modali ies o analyses and
obse a ions gi e aluable in o ma ion o u he planning o
in i o es ing o accines and immune modula o s in humanized
mice. The 3R p inciple (Replace animal es ing, Reduce he num-
be o animals, and Re ine he analyses) can hus be ad anced o
Reduce and Re ine: (i) by using (a leas ini ially) emale mice and
(ii) explo ing bio-in o ma ics me hods such as ANN o comple-
men adi ional s a is ical analyses in o de o de ine he mos
impo an issues (such as spleen and lymph nodes) and he PCA
ha e eal signa u es and co ela ions o immune esponses o
di e en lympha ic issues.
e hICs s A eMeN
All subjec s dona ing co d blood p o ided w i en in o med
consen . This s udy was app o ed by he E hics Commi ee o
Hanno e Medical School.
AU hoR CoN RIBU IoNs
RS planed he p ojec , designed expe imen s, ob ained unding
and egula o y app o als, en olled collabo a o s, in e p e ed he
da a, and w o e and edi ed he manusc ip . VV conduc ed expe i-
men s, analyzed da a, and w o e he i s manusc ip d a . BS, ST,
AS, LG, and CR assis ed in p epa a ion and analyses o human-
ized mice. CF pe o med he human cy okine a ay analyses. CK
assis ed in he p ocu emen and collec ion o HSC o he s udies.
LS pe o med he s a is ical analyses. AR, PR, HH, and MM-H
pe o med he ANN and PCA analyses, in e p e ed he da a,
and w o e and edi ed he manusc ip . SK and UK assis ed in he
execu ion o he iDCpp65 quali y con ol analyses, and e ised
he manusc ip .
ACKNoWLedGMeN s
The au ho s hank all o he cu en and pas membe s o he
Regene a i e Immune The apies Applied Labo a o y o hei
aluable con ibu ions. The au ho s hank Sebas ian Binde o
e ising he manusc ip .
FUNdING
This wo k was suppo ed by g an s o he Ge man Resea ch
Council (DFG/SFB738 P ojec A6 o RS; DFG/REBIRTH Uni
6.4 o RS, Uni 6.3 o CF) and he Ge man Cen e o In ec ions
Resea ch (DZIF-TTU07.803 o RS). VV ecei ed a DAAD/ZIB
Ph.D. ellowship, ST ecei ed a RegSci Ph.D. ellowship and CDR
ecei ed a CNPq “Sciences wi hou Bo de s” pos -doc o al el-
lowship. HH and AR would like o acknowledge he SYSMIFTA
ERACoSysMed g an (031L0085B) o he inancial suppo o
his wo k. MM-H and HH we e suppo ed by he Ge man Fede al
Minis y o Educa ion and Resea ch wi hin he Measu es o he
Es ablishmen o Sys ems Medicine, p ojec SYSIMIT (BMBF
eMed p ojec SYSIMIT, FKZ: 01ZX1308B and 01ZX1608B).
PR and MM-H we e suppo ed by he Human F on ie Science
P og am (RGP0033/2015).
sUPPLeMeN ARY MA eRIAL
The Supplemen a y Ma e ial o his a icle can be ound online a
h p://www. on ie sin.o g/a icle/10.3389/ immu.2017.01709/
ull#supplemen a y-ma e ial.
ReFeReNCes
1. Kenney LL, Shul z LD, G eine DL, B ehm MA. Humanized mouse models o
ansplan immunology. Am J T ansplan (2016) 16(2):389–97. doi:10.1111/
aj .13520
2. Theocha ides AP, Rong aux A, F i sch K, Fla ell RA, Manz MG. Humanized
hema o-lymphoid sys em mice. Haema ologica (2016) 101(1):5–19.
doi:10.3324/haema ol.2014.115212
3. Walsh NC, Kenney LL, Jangalwe S, A yee KE, G eine DL, B ehm MA,
e al. Humanized mouse models o clinical disease. Annu Re Pa hol (2017)
12:187–215. doi:10.1146/annu e -pa hol-052016-100332
4. B ehm MA, Cu hbe A, Yang C, Mille DM, Dilo io P, Laning J, e  al.
Pa ame e s o es ablishing humanized mouse models o s udy human
immuni y: analysis o human hema opoie ic s em cell eng a men in h ee
immunode icien s ains o mice bea ing he IL2 γnull mu a ion. Clin
Immunol (2010) 135:84–98. doi:10.1016/j.clim.2009.12.008
18
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
5. Ha is DT, Badowski M. Long e m human econs i u ion and immune aging
in NOD-Rag (-)-gamma chain (-) mice. Immunobiology (2014) 219(2):131–7.
doi:10.1016/j.imbio.2013.08.013
6. Audige A, Rocha MA, Li D, I ic S, Fah ny A, Mulle CKS, e  al. Long-
e m leukocy e econs i u ion in NSG mice ansplan ed wi h human co d
blood hema opoie ic s em and p ogeni o cells. BMC Immunol (2017)
18(1):28. doi:10.1186/s12865-017-0209-9
7. Hun ing on ND, Leg and N, Al es NL, Ja on B, Weije K, Ple A, e  al.
IL-15 ans-p esen a ion p omo es human NKcell de elopmen and di e -
en ia ion in i o. J Exp Med (2009) 206(1):25–34. doi:10.1084/jem.20082013
8. Rong aux A, Willinge T, Ma inek J, S owig T, Gea y SV, Teichmann LL,
e al. De elopmen and unc ion o human inna e immune cells in a human-
ized mouse model. Na Bio echnol (2014) 32(4):364–72. doi:10.1038/nb .2858
9. Danne R, Chaudha i SN, Rosenbe ge J, Su ls J, Richie TL, B umeanu TD,
e  al. Exp ession o HLA class II molecules in humanized NOD.Rag1KO.
IL2RgcKO mice is c i ical o de elopmen and unc ion o human T and
Bcells. PLoS One (2011) 6(5):e19826. doi:10.1371/jou nal.pone.0019826
10. Shul z LD, Sai o Y, Najima Y, Tanaka S, Ochi T, Tomizawa M, e al. Gene a ion
o unc ional human T-cell subse s wi h HLA- es ic ed immune esponses
in HLA class I exp essing NOD/SCID/IL2 γnull humanized mice. P oc Na l
Acad Sci U S A (2010) 107:13022–7. doi:10.1073/pnas.1000475107
11. T aggiai E, Chicha L, Mazzucchelli L, B onz L, Pi a e i JC, Lanza ecchia A,
e al. De elopmen o a human adap i e immune sys em in co d blood cell-
ansplan ed mice. Science (2004) 304(5667):104–7. doi:10.1126/science.
1093933
12. Tonomu a N, Habi o K, Shimizu A, Sykes M, Yang Y-G. B ie epo an igen-
speci ic human T-cell esponses and T cell – dependen p oduc ion o
human an ibodies in a humanized mouse model. Blood (2008) 111:4293–6.
doi:10.1182/blood-2007-11-121319.The
13. Lan P, Tonomu a N, Shimizu A, Wang S, Yang YG. Recons i u ion o a unc-
ional human immune sys em in immunode icien mice h ough combined
human e al hymus/li e and CD34+ cell ansplan a ion. Blood (2006)
108(2):487–92. doi:10.1182/blood-2005-11-4388
14. Melkus MW, Es es JD, Padge -Thomas A, Ga lin J, Den on PW, O hieno FA,
e  al. Humanized mice moun speci ic adap i e and inna e immune
esponses o EBV and TSST-1. Na Med (2006) 12(11):1316–22. doi:10.1038/
nm1431
15. Sai o Y, Ellegas JM, Manz MG. Gene a ion o humanized mice o
analysis o human dend i ic cells. Me hods Mol Biol (2016) 1423:309–20.
doi:10.1007/978-1-4939-3606-9_22
16. Bol KF, Sch eibel G, Ge i sen WR, de V ies IJ, Figdo CG. Dend i ic cell-
based immuno he apy: s a e o he a and beyond. Clin Cance Res (2016)
22(8):1897–906. doi:10.1158/1078-0432.CCR-15-1399
17. Lapen a C, San ini SM, Spada M, Dona i S, U bani F, Accapezza o D, e al.
IFN-alpha-condi ioned dend i ic cells a e highly e icien in inducing
c oss-p iming CD8(+) Tcells agains exogenous i al an igens. Eu J Immunol
(2006) 36(8):2046–60. doi:10.1002/eji.200535579
18. Meixlspe ge S, Leung CS, Rame PC, Pack M, Vanoaica LD, B e on G, e al.
CD141+ dend i ic cells p oduce p ominen amoun s o IFN-alpha a e
dsRNA ecogni ion and can be a ge ed ia DEC-205 in humanized mice.
Blood (2013) 121(25):5034–44. doi:10.1182/blood-2012-12-473413
19. Salgue o G, Daen hanasanmak A, Munz C, Rayko a A, Guzman CA, Riese P,
e al. Dend i ic cell-media ed immune humaniza ion o mice: implica ions
o allogeneic and xenogeneic s em cell ansplan a ion. J Immunol (2014)
192(10):4636–47. doi:10.4049/jimmunol.1302887
20. Daen hanasanmak A, Salgue o G, Sunda ase y BS, Waskow C, Cosgun KN,
Guzman CA, e al. Enginee ed dend i ic cells om co d blood and adul blood
accele a e e ec o Tcell immune econs i u ion agains HCMV. Mol The
Me hods Clin De (2015) 1:14060. doi:10.1038/m m.2014.60
21. Sunda ase y BS, Kloess S, Obe schmid O, Naundo S, Kuehlcke K,
Daen hanasanmak A, e al. Gene a ion o len i i us-induced dend i ic cells
unde GMP-complian condi ions o adap i e immune econs i u ion agains
cy omegalo i us a e s em cell ansplan a ion. J T ansl Med (2015) 13:240.
doi:10.1186/s12967-015-0599-5
22. Boppana SB, B i WJ. Recogni ion o human cy omegalo i us gene p oduc s
by HCMV-speci ic cy o oxic Tcells. Vi ology (1996) 222:293–6. doi:10.1006/
i o.1996.0424
23. Kla enbeek PL, Remme swaal EBM, en Be ge IJM, Doo ensplee ME,
an Schaik BDC, Es eld REE, e  al. Deep sequencing o an i i al T-cell
esponses o HCMV and EBV in humans e eals a s able epe oi e ha is
main ained o many yea s. PLoS Pa hog (2012) 8(9):e1002889. doi:10.1371/
jou nal.ppa .1002889
24. Sunda ase y B, Volk V, Theobald SJ, Ri inghausen S, Schaudien D,
Neuhaus V, e al. Human e ec o memo y T helpe cells engage wi h mouse
mac ophages and cause g a - e sus-hos -like pa hology in skin o human-
ized mice used in a nonclinical immuniza ion s udy. Am J Pa hol (2017)
187(6):1380–98. doi:10.1016/j.ajpa h.2017.02.015
25. Volk V, Schneide A, Spineli LM, G osshennig A, S ipecke R. The gende
gap: disc epan human T-cell econs i u ion a e co d blood s em cell ans-
plan a ion in humanized emale and male mice. Bone Ma ow T ansplan
(2016) 51(4):596–7. doi:10.1038/bm .2015.290
26. James G, Wi en D, Has ie T, Tibshi ani R. An In oduc ion o S a is ical
Lea ning. New Yo k: Sp inge -Ve lag (2013).
27. Koo eman NG, de Almeida PE, S ack JP, Nelakan i RV, Diecke S, Shao NY,
e  al. Alloimmune esponses o humanized mice o human plu ipo en
s em cell he apeu ics. Cell Rep (2017) 20(8):1978–90. doi:10.1016/j.cel ep.
2017.08.003
28. Haykin S. Neu al Ne wo ks: A Comp ehensi e Founda ion. NY: Macmillan
College Publishing (1994).
29. Ki schen RH, O’Higgins EA, Lee RT. The Royal London Space Planning:
an in eg a ion o space analysis and ea men planning: pa I: assessing
he space equi ed o mee ea men objec i es. Am J O hod Den o acial
O hop (2000) 118(4):448–55. doi:10.1067/mod.2000.109031
30. Sha ma B, Venugopalan K. Compa ison o neu al ne wo k aining unc ions
o hema oma classi ica ion in b ain CT images. IOSR J Compu Engin (2014)
16(1):31–5. doi:10.9790/0661-16123135
31. Jackson JE. A Use ’s Guide o P incipal Componen s. New Yo k: John Wiley &
Sons (2004). 569 p.
32. Salgue o G, Sunda ase y BS, Bo che s S, Wedekind D, Eiz-Vespe B,
Velaga S, e al. P econdi ioning he apy wi h len i i al ec o -p og ammed
dend i ic cells accele a es he homeos a ic expansion o an igen- eac i e
human Tcells in NOD.Rag1-/-.IL-2 gammac-/- mice. Hum Gene The (2011)
22(10):1209–24. doi:10.1089/hum.2010.215
33. Komandu i KV, John LSS, de Lima M, McMannis J, Rosinski S, McNiece I,
e  al. Delayed immune econs i u ion a e co d blood ansplan a ion is
cha ac e ized by impai ed hymopoiesis and la e memo y T-cell skewing.
Blood (2007) 110(13):4543–51. doi:10.1182/blood-2007-05-092130
34. Bosch M, Khan FM, S o ek J. Immune econs i u ion a e hema opoie ic
cell ansplan a ion. Cu Opin Hema ol (2012) 19(4):324–35. doi:10.1097/
MOH.0b013e328353bc7d
35. No a F, Doula o S, Dick JE. Eng a men o human hema opoie ic s em cells
is mo e e icien in emale NOD/SCID/IL-2Rgc-null ecipien s. Blood (2010)
115(18):3704–7. doi:10.1182/blood-2009-10-249326
36. Su he land JS, Spy oglou L, Mui head JL, Heng TS, P ie o-Hinojosa A,
P ince HM, e  al. Enhanced immune sys em egene a ion in humans
ollowing allogeneic o au ologous hemopoie ic s em cell ansplan a ion
by empo a y sex s e oid blockade. Clin Cance Res (2008) 14(4):1138–49.
doi:10.1158/1078-0432.CCR-07-1784
37. Gie ing-K oll C, Be ge P, Leppe dinge G, G ubeck-Loebens ein B. How
sex and age a ec immune esponses, suscep ibili y o in ec ions, and esponse
o accina ion. Aging Cell (2015) 14(3):309–21. doi:10.1111/acel.12326
38. Shou al R, Bondi O, Mishan H, Shimoni A, Unge R, Nagle A. Applica ion
o machine lea ning algo i hms o clinical p edic i e modeling: a da a-
mining app oach in SCT. Bone Ma ow T ansplan (2014) 49(3):332–7.
doi:10.1038/bm .2013.146
39. Shou al R, Labopin M, Unge R, Giebel S, Cice i F, Schmid C,
e  al. P edic ion o hema opoie ic s em cell ansplan a ion ela ed
mo ali y- lessons lea ned om he in-silico app oach: a Eu opean Socie y
o Blood and Ma ow T ansplan a ion Acu e Leukemia Wo king Pa y
Da a Mining S udy. PLoS One (2016) 11(3):e0150637. doi:10.1371/jou nal.
pone.0150637
40. Michael A, Ball G, Qua an N, Wushishi F, Russell N, Whelan J, e  al.
Delayed disease p og ession a e allogeneic cell accina ion in ho mone-
esis an p os a e cance and co ela ion wi h immunologic a iables. Clin
Cance Res (2005) 11(12):4469–78. doi:10.1158/1078-0432.CCR-04-2337
41. Mu phy GP, Snow P, Simmons SJ, Tjoa BA, Roge s MK, B and J,
e  al. Use o a i icial neu al ne wo ks in e alua ing p ognos ic ac o s
de e mining he esponse o dend i ic cells pulsed wi h PSMA pep ides
19
Volk e al. Lea ning Machine o Humanized Mice
F on ie s in Immunology | www. on ie sin.o g Decembe 2017 | Volume 8 | A icle 1709
in p os a e cance pa ien s. P os a e (2000) 42(1):67–72. doi:10.1002/
(Sici)1097-0045(20000101)42:1<67:Aid-P os8>3.0.Co;2-I
42. Milanez-Almeida P, Meye -He mann M, Toke A, Khailaie S, Huehn J. Foxp3+
egula o y T-cell homeos asis quan i a i ely di e s in mu ine pe iphe al
lymph nodes and spleen. Eu J Immunol (2015) 45(1):153–66. doi:10.1002/
eji.201444480
43. Ma ino S, El-Kebi M, Ki schne D. A hyb id mul i-compa men model o
g anuloma o ma ion and Tcell p iming in ube culosis. J Theo Biol (2011)
280(1):50–62. doi:10.1016/j.j bi.2011.03.022
Con lic o In e es S a emen : One o he co esponding au ho s is cu en ly
applying o a pa en ela ed o he con en o he manusc ip : R. S ipecke,
G. Salgue o, A. Daen hasanmak, A. Ganse . “Induced dend i ic cells and uses
he eo ” (PCT/EP2013/052485). All o he au ho s decla e ha he esea ch was
conduc ed in he absence o any comme cial o inancial ela ionships ha could
be cons ued as a po en ial con lic o in e es .
Copy igh © 2017 Volk, Reppas, Robe , Spineli, Sunda ase y, Theobald, Schneide ,
Ge asch, De es Ro h, Klöss, Koehl, Kaisenbe g, Figuei edo, Ha ziki ou, Meye -
He mann and S ipecke. This is an open-access a icle dis ibu ed unde he e ms
o he C ea i e Commons A ibu ion License (CC BY). The use, dis ibu ion o
ep oduc ion in o he o ums is pe mi ed, p o ided he o iginal au ho (s) o licenso
a e c edi ed and ha he o iginal publica ion in his jou nal is ci ed, in acco dance
wi h accep ed academic p ac ice. No use, dis ibu ion o ep oduc ion is pe mi ed
which does no comply wi h hese e ms.