Full Te ms & Condi ions o access and use can be ound a
h ps://www. and online.com/ac ion/jou nalIn o ma ion?jou nalCode=koni20
OncoImmunology
ISSN: (P in ) (Online) Jou nal homepage: h ps://www. and online.com/loi/koni20
Expanded NK cells om umbilical co d blood
and adul pe iphe al blood combined wi h
da a umumab a e e ec i e agains umo cells
om mul iple myeloma pa ien s
Chan al Reina-O iz , Michael Cons an inides , Alexis Fayd-He be-de-
Mauda e , Jessy P ésumey , Ja ie He nandez , Guillaume Ca on , Da id
Gi aldos , Rosana Díez , Isabel Izquie do , Gemma Azace a , Luis Palome a ,
Isabel Ma zo , Ja ie Na al , Albe o Anel & Ma ín Villalba
To ci e his a icle: Chan al Reina-O iz , Michael Cons an inides , Alexis Fayd-He be-de-
Mauda e , Jessy P ésumey , Ja ie He nandez , Guillaume Ca on , Da id Gi aldos , Rosana
Díez , Isabel Izquie do , Gemma Azace a , Luis Palome a , Isabel Ma zo , Ja ie Na al , Albe o
Anel & Ma ín Villalba (2021) Expanded NK cells om umbilical co d blood and adul pe iphe al
blood combined wi h da a umumab a e e ec i e agains umo cells om mul iple myeloma
pa ien s, OncoImmunology, 10:1, 1853314, DOI: 10.1080/2162402X.2020.1853314
To link o his a icle: h ps://doi.o g/10.1080/2162402X.2020.1853314
© 2020 The Au ho (s). Published wi h
license by Taylo & F ancis G oup, LLC. View supplemen a y ma e ial
Published online: 29 Dec 2020. Submi you a icle o his jou nal
A icle iews: 386 View ela ed a icles
View C ossma k da a
ORIGINAL RESEARCH
Expanded NK cells om umbilical co d blood and adul pe iphe al blood combined
wi h da a umumab a e e ec i e agains umo cells om mul iple myeloma pa ien s
Chan al Reina-O iz
a
, Michael Cons an inides
b
, Alexis Fayd-He be-de-Mauda e
b
, Jessy P ésumey
b
, Ja ie He nandez
b
,
Guillaume Ca on
c
, Da id Gi aldos
a
, Rosana Díez
d
, Isabel Izquie do
d
, Gemma Azace a
e
, Luis Palome a
e
, Isabel Ma zo
a
,
Ja ie Na al
a
, Albe o Anel
a
*, and Ma ín Villalba
b,
*
a
Apop osis, Immuni y & Cance G oup, Dep . Biochemis y and Molecula and Cell Biology, Facul y o Sciences, Campus San F ancisco Sq., Uni e si y
o Za agoza and A agón Heal h Resea ch Ins i u e (IIS A agón), Za agoza, Spain;
b
CHU Mon pellie , IRMB, Mon pellie , F ance;
c
Dépa emen
d’Héma ologie Clinique, CHU, Mon pellie , F ance;
d
Hema ology Depa men , Miguel Se e Hospi al, Za agoza, Spain;
e
Hema ology Depa men ,
Lozano Blesa Hospi al, Za agoza, Spain;
IRMB, Uni Mon pellie , INSERM, CHU Mon pellie , Mon pellie , F ance;
ABSTRACT
In his s udy we e alua ed he po en ial o expanded NK cells (eNKs) om wo sou ces combined wi h he mAbs
da a umumab and pemb olizumab o a ge p ima y mul iple myeloma (MM) cells ex i o. In o de o asce ain
he bes sou ce o NK cells, we expanded and ac i a ed NK cells om pe iphe al blood (PB) o heal hy adul
dono s and om umbilical co d blood (UCB). The esul ing expanded NK (eNK) cells exp ess CD16, necessa y o
ca ying ou an ibody-dependen cellula cy o oxici y (ADCC). Cy o oxici y assays we e pe o med on bone
ma ow aspi a es o 18 MM pa ien s and 4 pa ien s wi h monoclonal gammopa hy o unde e mined signi i-
cance (MGUS). Exp ession le els o PD-1 on eNKs and PD-L1 on MM and MGUS cells we e also quan i ied. Resul s
indica e ha mos eNKs ob ained using ou expansion p o ocol exp ess a low pe cen age o PD-1
+
cells. UCB
eNKs we e highly cy o oxic agains MM cells and addi ion o da a umumab o pemb olizumab did no u he
inc ease hei cy o oxici y. PB eNKs, while e ec i e agains MM cells, we e signi ican ly mo e cy o oxic when
combined wi h da a umumab. In a mino i y o cases, eNK cells showed a de ec able popula ion o PD1
+
cells.
This co ela ed wi h low cy o oxic ac i i y, pa icula ly in UCB eNKs. Addi ion o pemb olizumab did no es o e
hei ac i i y. Resul s indica e ha UCB eNKs a e o be p e e en ially used agains MM in he absence o
da a umumab while PB eNKs ha e signi ican cy o oxic ad an age when combined wi h his mAb.
ARTICLE HISTORY
Recei ed 24 July 2020
Re ised 11 No embe 2020
Accep ed 11 No embe 2020
KEYWORDS
Mul iple Myeloma; cell
The apy and
Immuno he apy; NK cells;
da a umumab; cy o oxici y;
PD-1
1. In oduc ion
The immune sys em la gely p e en s he de elopmen o umo s
bu clinical cance s e ade his immune su eillance. Mul iple
pa hways help c ea e an immunosupp essi e umo en i on-
men , in e e ing wi h umo an igen p esen a ion o cy o oxic
T cells.
1
Tumo cells hide he exp ession o umo -speci ic an i-
gens, limi ing he speci ic T cell an i- umo esponse.
2
NK cells,
which a e no an igen speci ic, p o ide a po en an i umo
esponse and a e a sou ce o powe ul immuno he apies.
3
NK
cells a e used in he clinic, especially in hema ological cance s
wi h poo p ognosis, and can sa ely be used in allogeneic
se ings.
4–6
Thei combina ion wi h an i- umo an ibodies,
h ough an ibody-dependen cellula cy o oxici y (ADCC),
o e s in e es ing he apeu ic oppo uni ies.
7,8
The immune checkpoin s e e o molecules ha inhibi o
modula e immune esponses, being CTLA-4 and PD-1 he wo
mos s udied, especially in he con ex o T cell ac i a ion.
9–11
The use o an i-PD1 blocking an ibodies such as pemb olizu-
mab and ni olumab has become a i s -line ea men in
umo s wi h poo p ognosis.
12,13
Some epo s indica e ha
PD-L1 exp ession in umo cells esul s in unc ional impai -
men o PD-1
+
NK cells.
14–16
While ce ain epo s show ha
NK cell unc ion can be pa ially es o ed using blocking
mAbs,
15,16
o he s show a mo e acu e impai men ha equi es
cy okines such as IL-2 and IL-15 o unc ional es o a ion.
14
Rega ding NK cells, addi ional molecules ac as checkpoin
inhibi o s, such as he inhibi o y NK cell ecep o NKG2A,
which inhibi s NK cell ac i i y when liga ed by HLA-E,
exp essed on he su ace o umo cells. The use o a NKG2A
blocking mAb, monalizumab, o e s p omise as a new umo
immuno he apy.
17
O he NK cell checkpoin mechanisms
include blocking he ac i a ing ecep o NKG2D by umo
shedding o hei MIC ligands,
18
o he nega i e ac ion o
LAG-3 and TIM-3.
19
Mul iple myeloma (MM) a ises om uncon olled p oli -
e a ion o abno mal plasma cells and accoun s o 10–20% o
all hema ological neoplasms and 0.9% o all newly diagnosed
cance cases wo ldwide.
20
MM is no mally p eceded by
a p emalignan phase, e med monoclonal gammopa hy o
unde e mined signi icance (MGUS). MGUS is ound in 3% o
CONTACT Albe o Anel [email p o ec ed] Apop osis, Immuni y & Cance G oup, Dep . Biochemis y and Molecula and Cell Biology, Facul y o Sciences, Campus
San F ancisco Sq., Uni e si y o Za agoza and A agón Heal h Resea ch Ins i u e (IIS A agón), Za agoza, 50009, Spain; Ma ín Villalba. INSERM U1183. Ins i u e o
Regene a i e Medicine and Bio he apy. 80, A . Augus in Fliche. 34295. Mon pellie Cedex 5. F ance
*Sha ed senio au ho ship
Supplemen al da a o his a icle can be accessed on he publishe ’s websi e.
ONCOIMMUNOLOGY
2020, VOL. 10, NO. 1, e1853314 (11 pages)
h ps://doi.o g/10.1080/2162402X.2020.1853314
© 2020 The Au ho (s). Published wi h license by Taylo & F ancis G oup, LLC.
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion-NonComme cial License (h p://c ea i ecommons.o g/licenses/by-nc/4.0/), which pe mi s
un es ic ed non-comme cial use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
he popula ion abo e he age o 50, bu he a e o p og ession
om MGUS o MM is e y low.
21
O e he pas wo decades,
ea men egimens and su i al a es o MM pa ien s ha e
wi nessed a adical imp o emen , wi h p o easome inhibi o s
and monoclonal an ibodies (mAbs) as he main con ibu o s.
22
Howe e , al hough o e all su i al and pa ien ou comes ha e
conside ably imp o ed, d ug esis ance is s ill a majo
conce n,
23
mainly in high- isk pa ien s, such as pa ien s wi h
(4;14) o del17p. Fo his eason, no el and mo e e icien
he apeu ic app oaches a e needed.
NK cell-based immuno he apy could ha e a posi i e impac
in MM ea men . One obs acle is he immunosupp essi e
umo mic oen i onmen p esen in MM pa ien s, in whom
almos all a ms o he immune sys em a e sub e ed.
24
Fo
example, NK cells om MM pa ien s a e ine ec i e agains
disease p og ession.
25
Hence, allogeneic ac i a ed and expanded
NK cells could be an in e es ing app oach. In ac , NK-based
ea men s ha e been ecen ly op imized o i s use in MM,
26–30
including a combina ion o a umo NK cell line wi h he an i-
CD38 mAb da a umumab.
31
CD38 is highly exp essed in abe an MM cells while i s
exp ession on no mal lymphoid cells, including plasma cells, is
ela i ely low. Da a umumab shows e icacy in elapsed and
e ac o y MM pa ien s ea ed wi h a leas wo p io lines o
he apy.
32–34
NK cell-media ed cy o oxici y seems o be one o
he main mechanisms o i s an i-MM ac i i y. Since pa ien
NK cell s a us is a iable, his could explain di e ence in
esponses among pa ien s.
35
Ce ain epo s indica e PD-L1 exp ession in MM cells which
co ela es wi h disease p og ession om he MGUS s age.
36,37
Howe e , clinical ials using PD-1 o PD-L1 blocke s, alone o
in combina ion wi h o he ea men s ha e ailed in MM.
38
The po en cy o oxici y o ac i a ed NK cells om PB
39–41
o UCB cells
7,42
agains a a ie y o malignancies has p e-
iously been published by ou labo a o ies. In one s udy, he
expansion o hese UCB NK cells was also epo ed.
7
In he
p esen wo k, we compa e he cy o oxic capaci y o PB and
UCB eNKs agains MM pa ien samples.
2. Resul s
2.1. Pe iphe al and umbilical co d blood NK expansion
NK cells we e expanded using wo di e en p o ocols. UCB
NK cells, due o hei need o bo h KIR and KAR signals o
each a ma u e pheno ype, we e cul u ed wi h PLH, an EBV-
ans o med HLA-I
+
B lymphoblas oid cell line which wo ks as
an accesso y cell wi h bo h equi ed signals.
7
Con e sely, PB
NKs a e ully ma u e and need only ac i a ing signals hus he
EBV+ HLA-I nega i e cell line 721.221 was employed.
43
As
a i s s ep, T cells and NKT cells we e deple ed om he
cul u es using an i-CD3 mAb, o a o NK cell expansion
(see Figu e 1(a), d 0). Each p o ocol equi ed di e en a ios
o accesso y cells, while IL-2 and IL-15 we e added a he same
concen a ions. UCB NKs we e ea ed wi h accesso y cells and
cy okines e e y 3 d and bene i ed om li le manipula ion. PB
Figu e 1. P og ession o NK pheno ypes h ough he expansion p o ocol. Flow cy ome y do plo s showing CD3 APC (y-axis) and CD56 PE Vio 770 (x-axis) s aining o PB
(a) and UCB (b) NKs om d 0 h ough d 20 o expansion p o ocol indica ed in ma e ials and me hods. His og ams depic he p og ession o PB (c) and UCB (d) NK cells
exp essing CD16 h oughou he 20-d expansion. Numbe s indica e he pe cen age o CD16
+
NK cells. Cells we e s ained wi h CD16 VioBlue o Pe CP, as indica ed. Lines
co ela e o es ing e e y 5 d.
e1853314-2 C. REINA-ORTIZ ET AL.
NKs we e sus ained o 5–6 d be o e cul u e enewal and
wi hs ood daily manipula ion seen in Figu e 1. By d 20, almos
all cells om bo h sou ces p esen in he cul u e we e
CD56
+
CD3
−
NK cells (see Figu e 1(a,b)). The pu i y o NK
cells was 94.28 ± 2.08% o UCB and 95.8 ± 1.46% as an a e age
(see Suppl. Figu e 1 and Suppl. Table I). The CD3
+
ac ion,
which was success ully deple ed a d 0, did no o e ake he
CD56
+
popula ion, allowing o success ul expansion o NK
cells (Figu e 1(a,b)). Beginning wi h 1E6 NKs in each expan-
sion, PB NKs eached an a e age o 240E6 cells by d 20 and
UCB NKs a e aged a 700- old expansion (Suppl. Fig 2).
The ini ial cha ac e iza ion o he pheno ype o NK cells
expanded using EBV- ans o med LCL as eede s has been
pe o med in p e ious s udies o ou labo a o ies. A de ailed
desc ip ion o all genes and miRNA exp essed upon
a 5-d ac i a ion is ound in.
44
The exp ession o speci ic ac i-
a ing and inhibi o y NK cell ecep o s was s udied ollowing
his 5-d ac i a ion p o ocol, showing inc eases in he pe cen-
age o NKp30
+
and especially NKp44
+
popula ions.
40
The
pheno ype o eNK cells ollowing hese expansion and ac i a-
ion p o ocols as compa ed o eshly isola ed NK cells has also
been s udied in p e ious wo k om ou g oups. eNK cells om
UCB show inc eased exp ession o ac i a ion ma ke s such as
CD69 and educed ha o CD45RA, becoming CD45RA
dim
cells, keeping high CD16 exp ession.
7
Rega ding eNK cells
ob ained om PB, he signi ican inc ease in he NKp44
+
popula ion p e iously desc ibed in Sanchez-Ma inez e al.
40
was clea ly con i med in ou la es s udy.
45
Consequen ly, we
ha e de e mined he exp ession o NKp44 ligands on he su -
ace o pa ien MM cells and ound ain le els o exp ession
(Suppl. Fig 3). Thei scan exp ession le el does no jus i y he
inc ease in cy o oxici y o eNKs. Ra he , his inc ease in cy o-
oxici y could mo e closely be ela ed o he p e iously demon-
s a ed inc ease in g anzyme B exp ession obse ed in
ac i a ed NK cells.
40
Rega ding CD16 exp ession, PB NKs p esen ed high le els
a d 0, which was main ained h oughou he expansion p o-
ocol (see a ep esen a i e expansion in Figu e 1(c) and
a summa y o da a on all dono s in Suppl. Table II). In gene al,
esh UCB NKs also showed CD16 exp ession, and his le el o
exp ession was also main ained du ing he expansions (see
a ep esen a i e expansion in Figu e 1(d) and a summa y o
da a on all dono s in Suppl. Table II). The main enance o
CD16 exp ession in he expanded NK cells is impo an in
o de o combine hem wi h he apeu ic an ibodies.
As he a e age pu i y o NK cells a d 20 a e aged 95%, we
did no p oceed wi h u he NK cell isola ion be o e he
cy o oxici y expe imen s. We es ima ed ha bo h ypes o
eNK cells could be used in cy o oxici y expe imen s om
d 10 on, hough expansion con inued un il d 20 o ob ain
a maximum numbe o eNK cells.
2.2. Limi ed a icide o eNKs in he p esence o
da a umumab
As NK cells can exp ess CD38, we ha e analyzed his poin in
ou expanded NK cells. We show in Figu e 2(a) he CD38
exp ession in he 10 expansions pe o med. All expansions
exp essed CD38 o a ying deg ees. While mos expansions
had CD38 exp ession in he majo i y o hei NK cell popula-
ion, a ew expansions om bo h PB and UCB exp essed CD38
in only a mino i y o he eNK cell popula ion. As da a umu-
mab binds o CD38, when CD38
+
NKs a e in he p esence o
da a umumab, c oss-linking can occu be ween NK cells
Figu e 2. A. CD38 exp ession de e mined by low cy ome y on d-20 expanded NK cells (eNK), ob ained om PB o om UCB; B, D-20 expanded NK cells (eNK) ob ained
om PB (le panel) o om UCB ( igh panel) we e le un ea ed (Con ol), o hey we e incuba ed o 4 h wi h 5 μg/ml o da a umumab and/o 10 µg/ml o
pemb olizumab, as indica ed. A e he incuba ions, eNK cell dea h was es ima ed by 7-ADD labeling. Ho izon al lines indica e he mean cell dea h in each expe imen al
condi ion.
ONCOIMMUNOLOGY e1853314-3
leading o cell killing ia ADCC.
46
The da a umumab-induced
a icide among NK cells could deple e he numbe o unc-
ional eNK cells a ailable o ac agains MM cells. To examine
he ex en o a icide in ou eNK cells, we cul u ed he eNKs
om PB o UCB wi h da a umumab and/o pemb olizumab,
as a seconda y con ol. Da a umumab exhibi ed a limi ed cy o-
oxic e ec , a e aging less han 5% o speci ic cell dea h (Figu e
2(b)). Highe le els o CD38 exp ession on NK cells did no
co ela e wi h inc eased a icide. Pemb olizumab did no
ha e a cy o oxic e ec alone o when combined wi h da a u-
mumab. We conclude ha da a umumab-induced a icide
would no signi ican ly impai eNK cy o oxic po en ial,
al hough his should be es ed in each eNK p epa a ion be o e
use.
2.3. PD-1 exp ession on eNK cells
Nex , we es ed PD-1 exp ession in he 5 PB and 5 UCB
eNK used in ou s udy. As shown in Figu e 3, mos eNK
cells ob ained (4 ou o 5 bo h in he case o PB and UCB),
showed a e y limi ed popula ion posi i e o PD-1 exp es-
sion: be ween 3.9% and 8.3% o he o al eNK popula ion in
he case o eNK de i ed om PB and be ween a 6.5% and
a 12.7% in eNK cells de i ed om UCB. Howe e , eNK
cells om PB2 and UCB1 showed he p esence o a de ined
and signi ican popula ion posi i e o PD-1 exp ession,
which accoun ed o 30.9% in PB2 and 23.13% in UCB1.
The MFI o he whole eNK popula ion in UCB1 was 4146
while he a e age MFI in he o he 4 UCBs was 1615. In
he PB expansions, he MFI alue o PB2 was 1481, while
he a e age o he o he 4 dono s was 766. The popula ion
o eNK cells posi i e o PD-1 exp ession was ne e highe
han 31%, emaining a mino subse o he whole eNK
popula ion. This esul is in ag eemen wi h he epo ed
low PD-1 exp ession in mos eNK cells ob ained om PB
in ou p e ious s udy.
45
2.4. eNK cy o oxici y assays agains MM and MGUS
pa ien samples
We i s ob ained bone ma ow aspi a es o MM and MGUS
om 22 pa ien s a di e en s ages o disease. Hal o he MM
pa ien s had unde gone a leas one p e ious ea men and
had expe ienced ecu ence. Pa ien 1 (MM1) unde wen six
p e ious ea men s and had again elapsed. MGUS pa ien s
we e moni o ed bu had ecei ed no ea men a he ime o
he bone ma ow biopsies. Thei clinical da a a e p esen ed in
Table 1. All samples we e ozen in liquid ni ogen a he
momen o ex ac ion and hawed be o e being used in he
cy o oxici y assays.
To analyze he he apeu ic po en ial o bo h PB and UCB
eNKs, we es ed hem agains he same pa ien samples in he
p esence o in he absence o he mAbs. In o de o be e
unde s and he e ec o PD-1 exp ession on he cy o oxic
abili y o eNKs, esul s we e s a i ied a ending o PD-1
exp ession in eNKs.
We i s ocused on esul s ob ained using PD-1 nega i e
eNK cells on MM pa ien samples (Figu e 4). Nei he da a u-
mumab no pemb olizumab, alone o in combina ion, had
a signi ican cy o oxic e ec on MM samples (Figu e 4(a)).
On a e age, UCB eNKs we e mo e cy o oxic on MM cells
han PB eNKs when used as lone ea men (Figu e 4(a,b)).
UCB eNKs cy o oxici y was no signi ican ly inc eased wi h he
addi ion o da a umumab. This was no due o low CD16
exp ession, which was obse ed in mo e han 80% o he
di e en UCB eNKs popula ion (see Figu e 1(d) and Suppl.
Figu e 3. A. Pa e n o PD-1 exp ession on all d-20 expanded NK cells (eNK) used in he s udy, ob ained om PB (uppe panels) o om UCB (lowe panels). B, Summa y
o he pe cen ages o PD-1
+
eNK cells ob ained om UCB o om PB, as indica ed.
e1853314-4 C. REINA-ORTIZ ET AL.

Table II). In con as , PB eNKs inc eased hei speci ic cy o-
oxici y when combined wi h da a umumab om 25% o 38%
on a e age (Figu e 4(b)), close o he le el ob ained by using
UCB eNKs alone. Finally, as expec ed, pemb olizumab had no
signi ican e ec on hese PD-1–nega i e eNKs (Figu e 4(a,b)).
Nei he da a umumab no pemb olizumab exhibi ed app e-
ciable cy o oxici y when used alone o in combina ion on
MGUS samples (Figu e 5(a)). When eNKs we e es ed on
MGUS cells, we obse ed again a signi ican ly highe cy o oxi-
ci y o UCB eNKs han PB eNKs on he same samples (Figu e 5
(a,b)). Cy o oxici y o UCB eNKs was sligh ly highe on MGUS
cells han on MM cells. Again, he combina ion wi h da a u-
mumab did no u he inc ease he high le el o UCB eNK
cy o oxici y. Al hough we did no ind any e ec o pemb oli-
zumab on MM cells, we obse ed a small a e age inc ease o
cy o oxici y o PB eNKs on MGUS samples, bu i was no
s a is ically signi ican . The inc ease in cy o oxici y when com-
bining PB eNKs wi h da a umumab, al hough obse ed, was
also no s a is ically signi ican . Pemb olizumab had no u he
e ec . Cy o oxici y o PB eNKs was also highe on MGUS han
on MM cells. These da a, ob ained wi h eNKs om bo h PB
and UCB, could indica e ha MGUS pa ien s may bene i om
eNK ea men p io o disease p og ession.
2.5. Cy o oxici y o eNK cells wi h a signi ican PD-1
+
popula ion
Figu e 6(a,c) show esul s ob ained wi h samples om he h ee
pa ien s ea ed wi h he only eNK ob ained om PB ha
con ained a signi ican PD-1
+
popula ion (PB2). Cu iously,
he eNK om PB2 showed a highe o e all a e age cy o oxici y
han ha obse ed using PD-1-nega i e PB eNKs, a ound 40%
o speci ic cell dea h (Figu e 6(c)). Howe e , he inc ease in
cy o oxici y when PB eNKs we e combined wi h da a umu-
mab, and p e iously obse ed in Figu e 4, was los .
Fo he ou pa ien samples ea ed wi h he only eNK
ob ained om UCB ha con ained a signi ican PD-1
+
popula-
ion (UCB1), we obse ed a d ama ic dec ease in cy o oxici y
when compa ed o esul s ob ained om PD-1-nega i e UCB
eNKs (Figu e 6(b,d)). The speci ic cell dea h a e age d opped
om 42% (see Figu e 4(b)) o 20%. This low cy o oxici y le el
was no due o esis ance o he pa ien samples hemsel es, as
cells om hese same pa ien s we e sensi i e o PD-1-nega i e
PB1 eNKs, as shown in Figu e 4. Addi ion o da a umumab did
no inc ease he cy o oxici y o hese eNKs, e en hough CD16
Table 1. MM and MGUS pa ien clinical da a. (A) Age and p e ious ea men da a
we e calcula ed based on he da e o pa ien bone ma ow biopsy. Fo MGUS
pa ien s, no ea men was gi en. The Du ie-Salmon sco e is no applicable (NA)
o MGUS pa ien s. PD-L1 exp ession is based on ou da a. Pa ien s wi h 0 p e ious
ea men s indica e newly diagnosed MM. (B) Numbe s o MM pa ien s ea ed
wi h each o he UCB o PB NK cell expansions. (C) Numbe s o MGUS pa ien s
ea ed wi h each o he UCB o PB NK cell expansions.
A
MM 18 (82%) MGUS 4 (18%)
Age
(a ime o sample)
Mean, yea s ( ange) 68 (45–83) 59 (49–77)
Gende
Male 12 (67%) 3 (75%)
Female 6 (33%) 1 (25%)
Du ie Salmon sco e
I 5 (28%) NA
II 4 (22%) NA
III 9 (50%) NA
P e ious ea men s
(a ime o sample)
0 9 (50%) NA
1 4 (22%) NA
2 2 (11%) NA
3+ 3 (17%) NA
PD-L1 Exp ession
+ (>50% cells+) 9 (50%) 0 (0%)
– (<50% cells+) 9 (50%) 4 (100%)
B
MM PTS UCB eNK PB eNK
1,2,3,4 1 1
8,9,10 2 2
11,13,14,15 3 3
16,17,18,19 4 4
20,21,22 5 5
C
MGUS PTS UCB eNK PB eNK
5 – 1
6,7 2 –
12 3 3
Figu e 4. eNK cy o oxici y assays on samples om MM pa ien s. Cells om MM pa ien s we e le un ea ed (MM), o hey ea ed wi h da a umumab a 5 μg/ml (d),
wi h pemb olizumab a 10 μg/ml (p), o wi h hei combina ion (D + P), o incuba ed wi h PD-1-nega i e UCB o PB eNKs o 4 h, as indica ed, in he absence o in he
p esence o he indica ed concen a ions o da a umumab, pemb olizumab, o hei combina ion. A e he incuba ions, cell dea h was es ima ed by 7-AAD labeling in
ga ed a ge cells (a), as indica ed in Ma e ial and Me hods. In (b) esul s a e shown as he pe cen age o speci ic cell dea h induced, a e sub ac ing basal cell dea h in
each sample. Ho izon al lines indica e he mean cell dea h in each expe imen al condi ion. One way ANOVA, ANOVA pos hoc Tukey’s analysis we e applied o mul iple
compa isons among a ious g oups. * indica es p alue less han 0.05 and conside ed o be signi ican .
ONCOIMMUNOLOGY e1853314-5
exp ession was high. This dec ease in cy o oxici y was no due
o he inhibi o y e ec o PD-1 because pemb olizumab did no
inc ease cy o oxici y on any o he ou pa ien s es ed.
In e es ingly, UCB1 was he only expansion ha s opped p o-
li e a ing a d 15 (see Suppl. Fig 2).
To comple e he in o ma ion on he cy o oxici y expe i-
men s, we include as Supplemen a y Figu e 4 he indi idual
cy o oxici y da a ob ained om he 22 pa ien samples, each
ea ed wi h he eNK cells delinea ed in Table 1(b,c).
2.6. Pa e n o PD-L1 exp ession in MM cells
In o de o be e unde s and he possible ole o he PD-1/PD-
L1 axis in he egula ion o eNK cy o oxici y, we analyzed he
PD-L1 le els o each MM sample (Figu e 7(a)). In mos o he
samples, a popula ion o PD-L1
+
cells could be de ec ed among
MM cells, 40% being he a e age alue. Only MM3 could be
conside ed nega i e o PD-L1 exp ession. Thus, we di ided
MM samples as PD-L1
high
o PD-L1
low
depending on whe he
Figu e 5. eNK cy o oxici y assays on samples om MGUS pa ien s. Cells om MGUS pa ien s we e le un ea ed (MGUS), o hey we e ea ed wi h da a umumab a
5 μg/ml (d), wi h pemb olizumab a 10 μg/ml (p), o wi h hei combina ion (D + P), o incuba ed wi h PD-1-nega i e UCB o PB eNKs o 4 h, as indica ed, in he absence
o in he p esence o he indica ed concen a ions o da a umumab, pemb olizumab, o hei combina ion. A e he incuba ions, cell dea h was es ima ed by 7-AAD
labeling in ga ed a ge cells (a), as indica ed in Ma e ial and Me hods. In (b) esul s a e shown as he pe cen age o speci ic cell dea h induced, a e sub ac ing basal
cell dea h in each sample. Ho izon al lines indica e he mean cell dea h in each expe imen al condi ion. One way ANOVA, ANOVA pos hoc Tukey’s analysis we e applied
o mul iple compa isons among a ious g oups. * indica es p alue less han 0.05 and conside ed o be signi ican .
Figu e 6. Cy o oxici y o eNK cells ha con ained a PD-1
+
popula ion. (a, c) Cells om he h ee indica ed MM pa ien s we e le un ea ed (MM), o hey we e ea ed
wi h da a umumab a 5 μg/ml (d), wi h pemb olizumab a 10 μg/ml (p), o wi h hei combina ion (D + P), o incuba ed wi h eNK cells om PB2 o 4 h, as indica ed, in
he absence o in he p esence o he indica ed concen a ions o da a umumab, pemb olizumab, o hei combina ion. (b, d) Cells om he ou indica ed MM pa ien s
we e le un ea ed (MM), o hey we e ea ed wi h da a umumab a 5 μg/ml (D), wi h pemb olizumab a 10 μg/ml (P), o wi h hei combina ion (D + P), o incuba ed
wi h eNK cells om UCB1 o 4 h, as indica ed, in he absence o in he p esence o he indica ed concen a ions o da a umumab, pemb olizumab, o hei combina ion
A e he incuba ions, cell dea h was es ima ed by 7-AAD labeling in ga ed a ge cells (a, b), as indica ed in Ma e ial and Me hods. In (c, d) esul s a e shown as he
pe cen age o speci ic cell dea h induced, a e sub ac ing basal cell dea h in each sample. Ho izon al lines indica e he mean cell dea h in each expe imen al condi ion.
e1853314-6 C. REINA-ORTIZ ET AL.
his popula ion accoun ed o mo e o less han 40%, espec-
i ely. Nine pa ien s ell in o each ca ego y, al hough wi hin
each g oup he e was high a iabili y in exp ession le els. The
a e age MFI alue in he pa ien s conside ed PD-L1
high
was
10819, while in he pa ien s conside ed PD-L1
low
was 5720.
The PD-L1 labeling o each indi idual MM and MGUS
pa ien is shown in Suppl. Fig 5. We con i m p e ious da a
indica ing ha MGUS cells a e nega i e o PD-L1 exp ession,
while mos MM samples showed PD-L1 exp ession o some
deg ee.
35
Taking in o accoun ha only ew NK cell expansions con-
ained a signi ican PD-1
+
popula ion (PB2 and UCB1), we
we e able o s udy he e ec o pai ing hese eNK cells wi h PD-
L1
high
MM cells only in one pa ien o each UCB and PB
eNKs. Pa ien 9 (MM9), in which 79% o he cells exp essed
PD-L1, was one o he pa ien s ea ed wi h he PB2 eNKs. The
cy o oxici y o PB2 eNKs on MM9 cells was in e media e
(Figu e 7(b), ed ba s). O he MM samples ea ed wi h hese
eNKs, MM9 exhibi ed one o he lowes le els o cy o oxici y
(see Figu e 6(a,b)). In his espec , he in e ac ion o PD-1 on
eNKs wi h PD-L1 on MM cells co ela es wi h a he low
cy o oxici y le els. A sligh inc ease in cy o oxici y was
obse ed when PB2 eNKs we e combined wi h da a umumab.
Howe e , he expec ed inc ease in cy o oxici y due o PD-1
blocking by pemb olizumab was no obse ed. MM9 cells we e
also ea ed wi h eNKs om he PD-1 nega i e UCB2.
A simila le el o cy o oxici y was obse ed, sligh ly inc eased
by he combina ion wi h da a umumab and una ec ed by
incuba ion wi h pemb olizumab (Figu e 7(b), blue ba s).
Pa ien 1 (MM1) showed 54% o cells exp essing PD-L1,
and was one o he pa ien s ea ed wi h he UCB1 eNKs. In
ag eemen wi h esul s ob ained om he o he pa ien s ea ed
wi h his UCB, cy o oxici y was ma kedly low wi h no mo e
han 10% o speci ic cell dea h induc ion (compa e Figu e 7(c),
ed ba s, wi h Figu e 6(c,d)). Nei he da a umumab no PD-1
blocking wi h pemb olizumab inc eased he cy o oxici y o
UCB1 eNKs on cells om his pa ien . O no e, he same esul
was obse ed using he same UCB eNKs agains MM cells om
pa ien s 2, 3 and 4 (Figu e 6(c,d)) ha we e PD-L1
low
o
nega i e (see Figu e 7(a)). MM1 cells did no exhibi an in in-
sic esis ance o NK cell cy o oxici y, since when ea ed wi h
eNKs om PB1, ha we e PD-1 nega i e, showed an in e -
media e le el o cy o oxici y, somewha inc eased by combina-
ion wi h da a umumab (Figu e 7(c), blue ba s).
Pemb olizumab was ac i e in he expe imen al condi ions
used in his s udy. When used in combina ion wi h wo PB
eNKs ha exhibi ed PD-1 exp ession in a signi ican pe cen-
age o cells, pemb olizumab was able o inc ease cy o oxici y
on he PD-L1
+
B-CLL cell line Mec-1 (Supplemen al Fig. 6).
When his assay was pe o med using ano he 5 PB eNKs ha
did no exhibi PD-1 exp ession, pemb olizumab was wi hou
e ec (da a no shown).
3. Discussion
The p esen wo k demons a es how he ac i a ion and expan-
sion o allogeneic NK cells gene a es eNK cells ha a e ac i e
agains MM cells. The ac i a ion and expansion we e achie ed
in UCB and PB Nks by using LCL lymphoblas s as eede s in
combina ion wi h IL-2 and IL-15. In bo h cases, he inal eNK
p oduc exhibi ed CD16 exp ession on an impo an ac ion
o hei popula ion (mo e han 80%), which would allow o
Figu e 7. (a) Exp ession o PDL-1 on he su ace o MM samples. The black line co esponds o he a e age alue and is used o di ide samples as PD-L1
high
and PD-L1
low
.
Only cells om MM3 can be conside ed nega i e o PD-L1 exp ession. (b, c) Cy o oxici y assays using eNK cells ha con ained a PD-1
+
popula ion on PD-L1
high
MM
samples. B, MM9, which is PD-L1
high
, was ea ed wi h PB2 eNK cells ( ed ba s) o wi h PD-1 nega i e UCB2 eNK (blue ba s). C, MM1, which is PD-L1
high
, was ea ed wi h
UCB1 eNKs ( ed ba s) o wi h PD-1 nega i e PB1 (blue ba s). Cell dea h was es ed by 7-AAD labeling, as indica ed in he legends o Figu es 4, 5 and 6. Resul s a e shown
as he pe cen age o speci ic cell dea h induced, a e sub ac ing basal cell dea h, which was ne e highe han 15%.
ONCOIMMUNOLOGY e1853314-7
hei combina ion wi h a a ie y o he apeu ic mAbs di ec ed
agains umo an igens.
7
This inc ease in cy o oxici y could be
associa ed wi h inc eased exp ession o ac i a ion ma ke s,
7
and o he ac i a ing ecep o NKp44.
45
Mo eo e , in
a simila s udy using a 5-d ac i a ion p o ocol, i was shown
ha he main change jus i ying he inc ease in cy o oxici y was
he ne inc ease in he le el o g anzyme B exp ession in
ac i a ed NK cells.
40
We obse ed how UCB eNKs showed a highe a e age cy o-
oxici y on MM cells han PB eNKs. The combina ion o PB
eNKs wi h he an i-CD38 mAb da a umumab, which is al eady
an e ec i e ea men in a ac ion o MM pa ien s, inc eased
hei cy o oxici y o he le els obse ed o UCB eNKs.
This is especially ele an as allogeneic NK cell ac i a ion-
based he apy, alone o in combina ion wi h an ibodies, is ye o
be app o ed o ea men o MM. While CAR T cell echnology is
a he o e on o cu en s udies, i is expensi e and has de i-
men al side e ec s, such as humo al immuni y inhibi ion ha
needs o be ea ed o e long pe iods o ime. eNK cell ea men s
do no gene a e a du able memo y esponse and would no ha e
his side e ec . Gi en he wide a ie y o p e ious ea men s
unde gone by he pa ien s in his s udy, elapsed and e ac o y
pa ien s would bene i om and espond o eNK cell he apy.
MM pa ien s p esen an immunosupp essi e umo mic o-
en i onmen , in which almos all a ms o he immune sys em
a e sub e ed.
22–24
Speci ically, NK cell ac i i y in MM pa ien s
is comp omised and pa ien NK cells a e ine ec i e agains he
de elopmen o he disease.
25
Howe e , clinical da a indica e
ha NK cells could be a he apeu ic s a egy in MM. In
a pionee ing wo k, Shi used haplo-iden ical misma ched NK
cells ea ed wi h IL-2 o 2 d (no expansion) esul ing in
posi i e ou comes in 5 ou o 10 elapsed MM pa ien s.
29
Ano he s udy used expanded NK cells om pa ien s as
a he apy, wi h esponse in 2 ou o 7 pa ien s.
30
A simila
s udy used expanded pa ien NK cells in combina ion wi h
an i-myeloma d ugs in i e elapsed MM pa ien s, showing
pa ial bu du able esponses.
26
Ne e heless, no s udies ha e
been pe o med using allogeneic expanded NK cells ob ained
om heal hy dono s. Mo e ecen ly, an inc ease in cy o oxici y
has been demons a ed on ex- i o cells ob ained om MM
pa ien s by combining expanded NK cells wi h ca ilzomib
27
and also by combining da a umumab wi h he ans o med NK
cell line KHYG1 ansien ly exp essing CD16.
31
Addi ionally,
p eclinical s udies ha e demons a ed he e iciency o an i-CS1
(SLAMF7, CD319) NK-CAR cells on MM cell lines.
47
A ecen
s udy using a humanized mice model demons a ed a posi i e
e ec o combining NK cells wi h he an i-CD137 mAb u elu-
mab, bu no wi h da a umumab.
48
PD-1 is an inhibi o y ecep o p esen on NK cells and
ac i a ed CD4
+
and CD8
+
T lymphocy es. I is in ol ed in
immunosupp ession by binding o i s ligands PD-L1 and PD-
L2, he o me showing a b oade exp ession.
9
MM is cha ac e -
ized by augmen ed PD-L1 exp ession
49,50
and some epo s
indica e ha PD-L1 exp ession co ela es wi h disease p og es-
sion om he MGUS s age.
36,37
Because o ha , se e al phase III
clinical ials we e conduc ed wi h checkpoin inhibi o s in
combina ion wi h es ablished MM ea men s:
51
pemb olizumab
plus lenalidomide and dexame hasone (Dex) (KEYNOTE-185,
NTC02579863); pemb olizumab plus pomalidomide (Poma)
and dexame hasone (KEYNOTE-183, NTC02576977); and
ano he s udy es ing h ee di e en combina ion egimens
(Poma and Dex s. ni olumab, Poma and Dex s. ni olumab,
elo uzumab, Poma and Dex; CheckMa e 602, NCT02726581).
Howe e , hese s udies we e discon inued due o an inc ease in
dea hs in he pemb olizumab g oup as well as no objec i e
esponses. Indeed, he FDA has pu on hold ano he clinical
ial combining he an i-PD-L1 mAb a ezolizumab wi h da a u-
mumab o simila easons (NCT02431208, see
38
).
Ou da a indica e ha mos eNK cells ob ained using he
p o ocol desc ibed showed a small popula ion o PD-1
+
cells,
sugges ing ha his inhibi o y pa hway would ha e a low ele-
ance in he ea men o MM pa ien s by hese eNKs. Howe e ,
in a mino i y o he cases, eNK cells showed a signi ican popula-
ion o PD1
+
cells. Pa icula ly in he case o UCB eNKs ha
con ained his PD-1
+
popula ion, cell dea h induc ion on MM
cells dec eased d ama ically and addi ion o pemb olizumab did
no es o e cy o oxici y. This esul was obse ed on MM cells
posi i e o nega i e o PD-L1 exp ession. This sugges s ha hese
eNK cells om UCB a e in insically non unc ional, independen
o he PD-1 signaling pa hway. This esul indica es ha he s a us
o PD-1 exp ession in he inal cellula UCB eNK p oduc should
be de e mined as a ma ke o lowe cy o oxic ac i i y.
The p esen da a could be suppo ed by s udies in 3D
cul u es o in in i o MM models. Howe e , 3D cul u es
equi e ela i ely long- e m pe iods in i o. Du ing his ime,
umo cells exp ess ligands ha we e no exp essed ex i o and
down-modula e he exp ession o o he molecules. Mo eo e ,
he e is a selec ion bias o ce ain clones ha g ow be e in
in i o cul u e. In con as , we examined umo cell sensi i i y
o eNK cells di ec ly ex i o, wi hou cul u ing hem in i o.
We chose o es umo cell sensi i i y o eNK cells quickly a e
ob aining he cells om he pa ien , a oiding cul u ing he
umo cells, a p ocess ha can a ec hei sensi i i y o allo-
geneic NK cells. Likewise, simila p oblems can be en isaged in
human MM cell eng a men in NSG mice, as he human
mic oen i onmen is no ep oduced.
UCB, a he han PB, con ains di e en NK cell p ogeni o
popula ions wi h he capaci y o di e en ia e in o NK cells
52
and his could o igina e NK cells wi h highe cy oly ic ac i i y
agains p ima y MM cells. In e es ingly, UCB NK cells, com-
pa ed o PB NK, ha e a highe exp ession o he bone ma ow
homing ecep o CXCR4
52
and somehow his could help o
ecognize p ima y MM cells ha niche in he bone ma ow.
The p esen da a indica e ha UCB eNKs a e o be p e e -
en ially used agains MM in he absence o da a umumab while
PB eNKs ha e signi ican cy o oxic ad an age when combined
wi h his mAb. Ou p esen da a, ob ained wi h MM pa ien s
anging om newly diagnosed o elapsed/ e ac o y, ein o ce
he easibili y o using allogeneic eNK cells in he ea men o
MM, mainly in iew o hei high cy o oxici y agains umo
cells in combina ion wi h da a umumab.
4. Ma e ials and Me hods
4.1. E hical s a emen
The use o human specimens o scien i ic pu poses was
app o ed by he F ench Na ional E hics Commi ee. All
e1853314-8 C. REINA-ORTIZ ET AL.