Multiplication of human natural killer cells by nanosized phosphonate-capped dendrimers

Article abstract of DOI:10.1002/anie.200604651

Non-natural born killers: The addition of phosphorus-containing dendrimers capped with phosphonate end groups to cultures of human peripheral blood mononuclear cells (white blood cells) strongly stimulates the selective multiplication of functional natural killer (NK) cells (see picture), which play a key role in anticancer immunity. Both the generation of the dendrimer and the type of end groups are important criteria for the activity. (Figure Presented).

Full text of DOI:10.1002/anie.200604651

Angewandte
Chemie
DOI: 10.1002/anie.200604651
Cell Multiplication
Multiplication of Human Natural Killer Cells by Nanosized
Phosphonate-Capped Dendrimers**
Laurent Griffe, Mary Poupot, Patrice Marchand, Alexandrine Maraval, CØdric-Olivier Turrin,
Olivier Rolland, Pascal MØtivier, GØrard Bacquet, Jean-Jacques FourniØ,
Anne-Marie Caminade,* RØmy Poupot,* and Jean-Pierre Majoral*
Dedicated to Professor Jean-Jacques Bonnet
Nanotechnology appears as the main innovative break-
through in the last decade for many scientific fields and has
in particular induced the emergence of “nanomedicine”.[1]
This rapidly expanding multidisciplinary field greatly benefits
from chemistry,with the elaboration of highly sophisticated
nanoscale and multifunctional devices.[2] Dendrimers,in
particular,which are precisely defined and tunable hyper-
branched synthetic polymers with layered architectures,[3]
show promise in several biomedical applications.[4] Indeed,
as a result of their size range (a few nanometers) and their
high structural and chemical homogeneity,they are suitable
for manipulations at the molecular level. Furthermore,their
multivalency owing to the presence of a large number of
terminal units may induce multiple simultaneous interactions
with biological entities that should be quantitatively and
qualitatively different from the individual interactions dis-
played by monomeric constituents (the “cluster glycoside
effect”[5]). One important field of research using dendrimers
in nanomedicine concerns oncology. Indeed,dendrimers can
be used as drug carriers that are able to enhance aqueous
solubility,improve drug transit across biological barriers,and
target the drug to diseased tissues.[4]
Besides the use of specific anticancer drugs,it is known
that the immune system is able to fight against cancers; the
challenge is to “manipulate” it to increase its efficiency.[6]
Peripheral blood immune cells are present within the
circulating pool of blood,easily accessible,and widespread
in the whole body,and thus they are a target of choice for such
purposes. The immune system in blood comprises several
kinds of cells derived from stem cells in bone marrow,in
particular natural killer (NK) cells,monocytes and dendritic
cells,which are part of the innate immunity,and B and T
lymphocytes which are part of the adaptive immunity (they
require a prior sensitization). Increasing artificially the
number of blood immune cells generally necessitates complex
and poorly available biological molecules/entities. However,
it has been shown previously that small molecules such as
pyrophosphates (referred to as phosphoantigens)[7] and
amino-bisphosphonates[8] can activate and/or multiply the
TCRgd+ subset of T lymphocytes,a group of T lymphocytes at
the borderline between adaptive and innate immunity.[9] Thus,
it appeared interesting to graft phosphorus derivatives and
especially phosphonates,which are more stable than phos-
phates and insensitive to phospholipases[10] and phospha-
tases,[11] as end groups of dendrimers and to test their behavior
towards human immune blood cells.
[*] Dr. L. Griffe,[+] Dr. P. Marchand, Dr. A. Maraval, Dr. C.-O. Turrin,
O. Rolland, Dr. A.-M. Caminade, Dr. J.-P. Majoral
Laboratoire de Chimie de Coordination, CNRS
205 route de Narbonne, 31077 Toulouse Cedex 4 (France)
Fax: (+33)5-6155-3003
E-mail: caminade@lcc-toulouse.fr
Up to now,very few dendrimers with phosphonate
(phosphonic acid salts) end groups have been synthesized
and characterized.[12,13] Here,we report several methods of
synthesizing dendrimers capped with phosphonate end
groups. The backbone of the selected dendrimers also
contains phosphorus as branching points,[14] and the biocom-
patibility of this backbone was previously demonstrated.[15]
The first series of phosphonate-capped dendrimers was
synthesized by nucleophilic substitution of the chlorine
atoms of dendrimer 1-Gn (n = 0–2)[16] by a phenol (a) derived
from tyramine bearing two dimethylphosphonate groups,in
the presence of cesium carbonate (Scheme 1). The azabisdi-
methylphosphonate end groups of dendrimers 2a-Gn were
then treated with trimethylsilyl bromide,methanol,and
finally with sodium hydroxide to afford dendrimers 3a-Gn.
Sodium salt derivatives were needed to ensure their solubility
in water (dendrimers with phosphonic acid end groups are
insoluble). Generations zero (six azabisphosphonate end
groups) to two (24 azabisphosphonate end groups) were
majoral@lcc-toulouse.fr
Dr. M. Poupot,[+] Dr. J.-J. FourniØ, Dr. R. Poupot
INSERM U.563
Centre de Physiopathologie de Toulouse-Purpan
Hôpital Purpan, 31300 Toulouse (France)
Fax: (+33)5-6274-4548
E-mail: remy.poupot@toulouse.inserm.fr
Dr. P. MØtivier, Dr. G. Bacquet
Centre de Recherches d’Aubervilliers, Rhodia
52 rue de la Haie Coq, 93308 Aubervilliers Cedex (France)
[+] Contributed equally to this work.
[**] We thank Rhodia (grants to L.G., P.M., and O.R.), the CNRS (grant
to A.M.), the EC (F.S.E. grant to M.P.), and the RØgion Midi-
PyrØnØes (“Biotherapy” program) for financial support. We also
thank Dr. Yannick Coppel for NMR spectra.
Supporting information for this article (including syntheses,
characterizations, copy of 1H, 13C, and 31P NMR spectra of all
compounds, materials and methods for biology, and flow cytometry
figures) is available on the WWW under http://www.angewand-
te.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 2523 –2526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2523

Communications
bond,thus we designed another way to synthesize 2b-G1
consisting of a condensation reaction of the aldehyde end
[16]
groups of 4-G1
with benzylamine to afford 5-G1. The
resulting imine bonds were then submitted to addition
reactions with dimethylphosphite to afford cleanly 2b-G1.
Dendrimers 2b-G1 and 2c-G1 were then converted into the
corresponding water-soluble phosphonic acid salts 3b-G1 and
3c-G1,respectively,by using the method already described in
Scheme 1. Note that 3b-G1 is also obtainable from 5-G1 by
addition of P(OSiMe)3 to the imines to afford 6-G1,followed
by reaction with MeOH and NaOH (Scheme 2).
The series of dendrimers 3d-Gn (n = 0–2)[17] with another
type of negatively charged end groups (carboxylates) was also
synthesized by carrying out a modified Doebner reaction on
dendrimers 4-Gn.
Scheme 1. Synthesis of first-generation dendrimer 3a-G1, and chemical
structures of zero- and second-generation dendrimers 3a-G0 and 3a-
G2, respectively.
synthesized to determine the influence of the loading in
phosphonate groups on the biological properties.
In a first attempt,the first generation of the symmetrical
azabisphosphonate dendrimers 3a-G1 was added to cultures
of human peripheral blood mononuclear cells (PBMCs; white
blood cells issued from healthy donors) supplemented with
interleukin-2 (IL-2). We recently reported that such den-
drimers are able to activate monocytes within 30 minutes but
not to multiply them.[18] To our great surprise,a very different
behavior was observed for longer times of culture. Indeed,an
important increase in the number of PBMCs was observed
(proliferation index 5.5 in two-week-old cultures with 20 mm
3a-G1) and phenotyping of the cells multiplied in cultures
with 3a-G1 revealed the prominence of NK cells (with some T
cells). Interestingly,these NK cells express a high level of the
membranous marker CD56 and a low level of the membra-
nous marker CD16,thus indicating that they are proliferative
NK cells[19] (see the Supporting Information). Experiments
with PBMCs obtained from six healthy donors revealed in all
cases an important increase in both the percentage and the
We also varied the surface of these dendrimers to
determine the influence of the type of end groups. We chose
in particular to graft an unsymmetrical bisphosphonate (to
see the influence of the local geometry),a monophosphonate
(to see the difference between monodentate and bidentate
end groups),and a carboxylate (to see if phosphonates are
needed or only negative charges). First,we synthesized two
phenols bearing either one (b) or two (c) dimethylphospho-
nates which were obtained by hydrophosphinylation of an
imine with dimethylphosphite,followed by a Kabachnik–
Fields reaction with dimethylphosphite and formaldehyde
(Scheme 2). The grafting of these phenols to the first-
generation dendrimer 1-G1 affords dendrimers 2b-G1 (aza-
monodimethylphosphonate end groups) and 2c-G1 (unsym-
metrical azabisdimethylphosphonate end groups). In the case
of the monodimethylphosphonate,we observed in some
experiments a small percentage of coupling through the NH
Scheme 2. Synthesis of dendrimers with monophosphonate (3b-G1) or unsymmetrical bisphosphonate (3c-G1) end groups.
www.angewandte.org ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2007, 46, 2523 –2526
2524

Angewandte
Chemie
number of NK cells when the results were compared with IL-2
alone and with IL-2 and 3a-G1 (Figure 1). After four weeks in
culture,a mean multiplication of the number of NK cells by a
factor of 105 was achieved in medium supplemented with IL-2
and 3a-G1 versus a mean multiplication only by a factor of 7.5
in medium supplemented with IL-2 alone. These large-scale
prototype cultures started with PBMCs comprising 1 million
NK cells on average; multiplications over 500-fold were
obtained with some donors (Figure 1).
Figure 2. Specific lysis (destruction) of various leukemia (A) and
carcinoma (B) cells by NK cells generated in cultures with dendrimer
3a-G1. Two ratios of NK/cancerous cells were used (white bars: 3:1;
gray bars: 10:1).
Figure 1. A) Number and B) percentage of NK cells from four-week-old
cultures with IL-2 alone (gray dots) or with IL-2 and 20 mm 3a-G1
dendrimer (black dots). Data points a–f represent six healthy donors.
Human NK cells constitute a key population of innate
immunity implicated in early immune responses against
viral,[20] bacterial,[21] and parasitic infections,[22] and they play
a crucial role in anticancer immunity.[23] However,the
proliferation of NK cells is extremely tedious to achieve[24]
and there are still no means of producing large batches of NK
cells necessary for the development of NK-cell-based thera-
pies. These results prompted us to verify if the NK cells
generated in cultures with 3a-G1 are fully functional and
mature. For this purpose,their ability relative to uncultured
NK cells to kill the same cancer cell lines was successfully
tested and found to show the same efficiency (15 cell lines
tested,seven different leukemia and eight different carci-
noma; see Figure 2). Furthermore and most importantly,
autologous lymphocytes were preserved (no aggressiveness of
the NK cells generated with dendrimers against lymphocytes
of the same donor).
These very important results encouraged us to determine
the main determinants in the properties of 3a-G1. First,we
checked the dose/effect relationship,which indicated that 3a-
G1 is the most active at 20 mm (Figure 3A). Then,we
established low-scale routine cultures (i.e. three-week cul-
tures starting with 1 ” 105 NK cells on average) to screen the
bioactivity of the variety of dendrimers described above. This
culture procedure was applied to determine if the activity is
only linked to the concentration of bisphosphonates in the
culture medium. Figure 3B demonstrates that the generation
has a greater influence than the total concentration of
bisphosphonates,with 3a-G1 being the most active of the
3a-Gn series.
Figure 3. A) Dose/effect titration for 3a-G1 used in large-scale cultures
(20 mL). B) Low-scale cultures (5 mL) with various concentrations (in
brackets) of 3a-Gn (n=0–2) or without dendrimer. The x axis is the
concentration of end groups (bisphosphonates) obtained by multi-
plying the molar concentration of dendrimer by 6 for 3a-G0, by 12 for
3a-G1, and by 24 for 3a-G2.
percentage of NK cells than unsymmetrical azabisphospho-
nate (3c-G1) or azamonophosphonate (3b-G1) groups when
grafted on first-generation dendrimers (Figure 4). Thus,
besides the generation,the geometry of the end groups
plays an important role also.
In conclusion,we have demonstrated that phosphonates
grafted to the surface of phosphorus-containing dendrimers
display the unexpected property of dramatically and selec-
tively promoting the multiplication of human natural killer
cells. Depending on the size of the dendrimers and on the
type,number,and even the geometry of the end groups they
bear,large differences in their bioactivity toward NK cell
multiplication are observed. In view of the virtually unlimited
number of chemical structures of dendrimers,a more
detailed,quantitative structure–activity relationship study is
needed. However,the multiplication observed of up to 500-
fold in certain cases is unprecedented. Furthermore,the
Comparison of the 3a-Gn series with the 3d-Gn series (n =
0–2) clearly indicated that azabisphosphonate end groups
rather than carboxylate end groups afford the highest multi-
plication of NK cells. Furthermore,symmetrical azabi-
sphosphonate groups gave better results concerning the
number of NK cells but similar results concerning the
Figure 4. Numbers (black bars) and percentages (gray bars) of NK
cells from three-week-old cultures with IL-2 alone or with IL-2 and
20 mm 3x-Gn (x=a–d; n=0–2) dendrimers. All experiments were
conducted on samples of blood issued from the same donor.
Angew. Chem. Int. Ed. 2007, 46, 2523 –2526
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2525

Communications
bioactivity of the NK cells generated in the presence of
Keywords: biological activity · dendrimers · immunology ·
NK cells · phosphonates
.
dendrimers is not modified,in contrast to what was previously
observed with known ligands activating NK cell receptors.[25]
Cultures with these dendrimers did not induce activation or
inhibition of the NK cells lytic response nor compromise
direct toxicity for their target cells and preserve autologous
lymphocytes. Thus,these dendrimers constitute a new tool in
“nanomedicine”. Taking into account the known properties of
NK cells,the straightforward production of therapeutic-scale
batches of functional NK cells might accelerate the advent of
NK cell therapies against infectious and malignant diseases.
[1] V. Wagner,A. Dullaart,A. K. Bock,A. Zweck, Nat. Biotechnol.
2006, 24,1211 – 1217.
[2] R. Haag,F. Kratz, Angew. Chem. 2006, 118,1218 – 1237; Angew.
Chem. Int. Ed. 2006, 45,1198 – 1215.
[3] a) G. R. Newkome,C. N. Moorefield,F. Vögtle, Dendrimers and
Dendrons: Concepts, Syntheses, Applications,Wiley-VCH,
Weinheim, 2001; b) J. M. J. FrØchet,D. A. Tomalia, Dendrimers
and Other Dendritic Polymers,Wiley,Chichester, 2001.
[4] C. C. Lee,J.A. MacKay,J. M. J. FrØchet,F. C. Szoka,
Biotechnol. 2005, 23,1517 – 1526.
[5] J. J Lundquist,E. J. Toone, Chem. Rev. 2002, 102,555 – 578.
[6] O. J. Finn, Nat. Rev. Immunol. 2003, 3,630 – 641.
[7] M. Poupot,J. J. FourniØ, Immunol. Lett. 2004, 95,129 – 138.
[8] V. Kunzmann,E. Bauer,J. Feurle,F. Weissinger,H. P. Tony,M.
Wilhelm, Blood 2000, 96,384 – 392.
Nat.
Experimental Section
See the Supporting Information for full experimental data,including
1
copies of H, 13C,and 31P NMR spectra and details of materials and
exhaustive methods for the biology part. The synthesis of dendrimer
3a-G1 is shown here as an example.
Synthesis of phenol a: A solution of formaldehyde (37% in water,
8 mL,104 mmol) was added to a solution of tyramine (7 g,51.2 mmol)
in THF (50 mL). The resulting mixture was stirred for 30 minutes at
room temperature,and then dimethylphosphite (10 mL,110 mmol)
was added. The mixture was stirred at room temperature for 24 h.
Brine (50 mL) was added to the crude material,which was then
extracted with ethyl acetate (3 ” 200 mL). The combined organic
phases were dried over MgSO4 and concentrated under reduced
pressure. The resulting crude oil was purified by column chromatog-
raphy (silica,acetone, Rf = 0.25) to afford phenol a as a colorless oil
(65% yield).
[9] W. K. Born,C. L. Reardon,R. L. OꢀBrien, Curr. Opin. Immunol.
2006, 18,31 – 38.
[10] W. Xu,S. A. Lee,T. G. Kutateladze,D. Sbrissa,A. Shisheva,
G. D. Prestwich, J. Am. Chem. Soc. 2006, 128,885 – 897.
[11] C. Belmant,E. Espinosa,F. Halary,Y. Tang,M.A. Peyrat,H.
Sicard,A. Kozikowski,R. Buelow,R. Poupot,M. Bonneville,
J. J. FourniØ, FASEB J. 2000,1669 – 1670.
[12] H. T. Chen,M. F. Neerman,A. R. Parrish,E. E. Simanek, J. Am.
Chem. Soc. 2004, 126,10044 – 10048.
[13] G. Bansal,J. E. I. Wright,C. Kucharski,H. Uludag,
Angew.
Synthesis of dendrimer 2a-G1: Cesium carbonate (428 mg,
1.314 mmol) and phenol a (501 mg,1.314 mmol) were added to a
solution of dendrimer 1-G1 (182 mg,9.95 ” 10 À2 mmol) in acetone
(5 mL). The reaction mixture was stirred at room temperature
overnight then centrifuged,and the resulting clear solution was
concentrated under reduced pressure. The residual oil was eluted on a
plug of silica with acetone to remove the unreacted phenol with
acetone/methanol (1:1) or acetone/water (7:3). The resulting solution
of dendrimer was concentrated to dryness under reduced pressure,
the residue was dissolved in CH2Cl2 (10 mL),and the solution was
dried over Na2SO4,filtered (micropore,0.2 mm),and finally concen-
trated under reduced pressure to afford dendrimer 2a-G1 as a
colorless oil (85% yield).
Chem. 2005, 117,3776 – 3780; Angew. Chem. Int. Ed. 2005, 44,
3710 – 3714.
[14] N. Launay,A.-M. Caminade,R. Lahana,J.-P. Majoral, Angew.
Chem. 1994, 106,1682 – 1684; Angew. Chem. Int. Ed. Engl. 1994,
33,1589 – 1592.
[15] a) T. R. Krishna,M. Parent,M. H. V. Werts,L. Moreaux,S.
Gmouh,S. Charpak,A.-M. Caminade,J.-P. Majoral,M. Blan-
chard-Desce, Angew. Chem. 2006, 118,4761 – 4764; Angew.
Chem. Int. Ed. 2006, 45,4645 – 4648; b) M. Blanzat,C. O. Turrin,
A. M. Aubertin,C. Vidal,A.-M. Caminade,J.-P. Majoral,I.
Rico-Lattes,A. Lattes, ChemBioChem 2005, 6,2207 – 2213.
[16] N. Launay,A.-M. Caminade,J.-P. Majoral, J. Organomet. Chem.
1997, 529,51 – 58.
Synthesis of dendrimer 3a-G1: Trimethylsilyl bromide (375 mL,
2.84 mmol) was added dropwise to a solution of dendrimer 2a-G1
(300 mg,5.03 ” 10 À2 mmol) in acetonitrile (15 mL) maintained at 08C.
The reaction mixture was stirred at room temperature overnight,then
the solvent was evaporated to dryness under reduced pressure. The
crude residue was washed with methanol (2 ” 5 mL) for one hour at
room temperature and dried under reduced pressure. The resulting
white solid was washed once with diethyl ether (20 mL) and then
transformed into its sodium salt by adding one equivalent of sodium
hydroxide per terminal phosphonic acid to a suspension of the
dendrimer in water (1 mL/100 mg). The resulting solution was
lyophilized to afford dendrimer 3a-G1 as a white powder (80% yield).
Biological procedures: Fresh blood samples were collected from
healthy adult donors,and PBMCs were prepared on a Ficoll-Paque
gradient by centrifugation (800g,30 min at RT). For NK cell
multiplication,PBMCs were cultured in complete RMPI 1640
medium supplemented with penicillin and streptomycin (both at
100 UmLÀ1),1 m m sodium pyruvate,10% heat-inactivated fetal calf
serum,and recombinant human IL-2 (400 UmL À1). Cultures were
started with 1.5 ” 106 cellsmLÀ1,and this density was maintained
along the culture by adding fresh medium. Sterile filtered solutions of
the specified dendrimers were added at a final concentration of 20 mm.
[17] M. K. Boggiano,G. J. A. A. Soler-Illia,L. Rozes,C. Sanchez,
C. O. Turrin,A.-M. Caminade,J.-P. Majoral, Angew. Chem. 2000,
112,4419 – 4424; Angew. Chem. Int. Ed. 2000, 39,4249 – 4254.
[18] M. Poupot,L. Griffe,P. Marchand,A. Maraval,O. Rolland,L.
Martinet,F. E. LꢀFaqihi-Olive,C. O. Turrin,A.-M. Caminade,
J. J. FourniØ,J.-P. Majoral,R. Poupot, FASEB J. 2006, 20,2339 –
2351.
[19] M. A. Cooper,T. A. Fehninger,M.A. Caligiuri, Trends Immu-
nol. 2001, 22,633 – 640.
[20] C. A. Biron,K. B. Nguyen,G. C. Pien,L. P. Cousens,T. P.
Salazar-Mather, Annu. Rev. Immunol. 1999, 17,189 – 220.
[21] P. Garcia-Penarrubia,F. T. Koster,R. O. Kelley,T. D. McDowell,
A. D. Bankhurst, J. Exp. Med. 1989, 169,99 – 113.
[22] T. Scharton-Kersten,A. Sher, Curr. Opin. Immunol. 1997, 9,44 –
51.
[23] a) J. N. Blattman,P. D. Greenberg, Science 2004, 305,200 – 205;
b) A. Cerwenka,L. L. Lanier, Nat. Rev. Immunol. 2001, 1,41 –
49.
[24] H. G. Klingemann, Cytotherapy 2005, 7,16 – 22.
[25] A. Moretta,C. Bottino,M. Vitale,D. Pende,C. Cantoni,M. C.
Mingari,R. Biassoni,L. Moreta, Annu. Rev. Immunol. 2001, 19,
197 – 223.
Received: November 15,2006
Published online: February 15,2007
2526
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2523 –2526