Science - USA (2019-01-18)

SCIENCE sciencemag.org

GRAPHIC: N. DESAI/

SCIENCE

after BMT. They propose that passive trans-

fer of antibodies that are matched to the la-

tent CMV strain in HSCT recipients might

constitute a powerful and easy therapeutic

approach to prevent CMV disease.

It is thought that ab T cells, gd T cells,

natural killer (NK) cells, and B cells all con-

tribute to curbing CMV infection ( 6 ) and

that anti-CMV CD8+ T cells are especially

important in CMV immune surveillance and

prevention of reactivation and dissemina-

tion. Analysis of cadaveric organ donor tis-

sue has revealed potential sites of active T

cell–mediated CMV clearance in the blood,

bone marrow, and spleen and reservoirs of

viral persistence in the lung coexisting with

active antiviral T cells ( 7 ). Indeed, transfer of

anti-CMV CD8+ T cells from donors is one

of the approaches investigated to treat CMV

reactivation in immunosuppressed patients

( 8 ). Martins et al. found that in mice, T cell

immunosuppression or depletion alone is

insufficient to enable CMV reactivation after

BMT, and only when host B cells and plasma

cells are eliminated [by graft-versus-host

disease (GVHD), a process by which donor

T cells can attack host cells, or by genetic

deficiency] in addition to T cell and NK cell

depletion does infection emerge (see the fig-

ure). Moreover, they demonstrate that pro-

tection from CMV reactivation conferred

by antibodies is much more efficient if it is

strain specific. Additionally, protection by

antibodies that prevent cell-to-cell dissemi-

nation is more effective than by antibod-

ies that prevent cell entry. Whether these

discoveries in mice—by using a different

conditioning regimen from that used in the

clinic and, by necessity, mouse rather than

human CMV (which are quite different)—

are relevant to the clinical situation merits

careful examination.

Recipients of allogeneic HSCT can be

treated with a variety of regimens to pre-

vent the host immune system from rejecting

the allogeneic stem cells, with a heavy focus

on T cell immunosuppression. These con-

ditioning regimens may include total-body

irradiation at high or low doses, cytoreduc-

ing chemotherapy drugs, and antithymocyte

globulin (ATG) or anti-CD52 (alemtuzumab)

to eliminate host T cells. Given the findings

of Martins et al., it is compelling to con-

sider how these clinical treatments may af-

fect antibody-producing B cells and plasma

cells because CMV reactivation in the clinic,

unlike in mice, can occur in the absence of

GVHD. Plasma cells are thought to be some-

what radioresistant, but ATG is a mixture

of polyclonal antibodies, some of which can

target B cells and plasma cells ( 9 ), and CD52

is expressed on both T and B cells. HSCT

recipients have been reported to experience

prolonged B cell depletion ( 10 ) and might

thus develop a reduction in CMV-specific

antibody titers after HSCT. Prospective fol-

lowing of CMV-specific antibody titers may

be of interest in the clinic to see whether this

is indeed the case.

Prevention or treatment of CMV with im-

munoglobulins has usually been ineffective

in HSCT recipients ( 11 ). The authors argue

that this may be because these reagents,

which come from sera pooled from many

donors, may not contain sufficient concen-

trations of antibodies specific to the CMV

strain infecting a given recipient. That hu-

mans can experience successive infections

with different CMV strains ( 12 ) supports

the possibility that the protective effect of

CMV antibodies is strain specific or at least

that antibodies generated after a given CMV

infection are not broadly cross-protective

against other strains. This is consistent with

the observation of Martins et al. that in their

mouse model, protection against CMV reac-

tivation is only conferred by strain-specific

serum and not when this serum is diluted

with sera containing antibodies reactive

to other strains. If confirmed in the clinic,

generation of a collection of sera specific for

each CMV genotype may be useful for pas-

sive antibody therapy in the future.

An additional point to consider is how

antibodies may work to prevent reactiva-

tion of CMV that is inside infected cells. The

authors show that neutralizing antibodies

that prevent cell infection are not as pro-

tective when injected in vivo as those that

prevent cell-to-cell dissemination in vitro,

which may provide insights for vaccine de-

sign. Much effort has been devoted to de-

veloping vaccines against human CMV ( 13 ).

The study of Martins et al. support giving

further consideration to the role of humoral

immunity in the prevention of reactivation

and dissemination of CMV after clinical

HSCT, the therapeutic potential of strain-

specific serum transfer, and the rational

vaccine design either for CMV-seronegative

recipients or to boost or broaden responses

of seropositive recipients before HSCT. j

REFERENCES

1. H. Kaminski, J. A. Fishman, Am. J. Transplant. 16 , 2254
   (2016).


2. D. Furman et al., Sci. Transl. Med. 7 , eaaa2293 (2015).


3. D. Lilleri, G. Gerna, Immunotherapy 8 , 1135 (2016).


4. E. Maffini et al., Expert Rev. Hematol. 9 , 585 (2016).


5. J. P. Martins et al., Science 363 , 288 (2019).


6. A. Huygens et al., Front. Immunol. 5 , 552 (2014).


7. C. L. Gordon et al., J. Exp. Med. 214 , 651 (2017).


8. M. Cobbold et al., J. Exp. Med. 202 , 379 (2005).


9. M. S. Zand et al., Transplantation 79 , 1507 (2005).


10. E. Corre et al., Haematologica 95 , 1025 (2010).


11. P. Raanani et al., J. Clin. Oncol. 27 , 770 (2009).


12. C. Smith et al., J. Virol. 90 , 7497 (2016).


13. P. R. Krause et al., Va c c i n e 32 , 4 (2013).

10.1126/science.aav9867

Host infected

fbroblast

CMV antibodies

CMV

dissemination

Host NK cell

Prevention of

CMV disease

Host anti-CMV

CD8+ T cell

Prevention of CMV dissemination

Prevention of CMV dissemination

Decreased CMV

antibodies

GVHD

Prevention of CMV dissemination

Host anti-CMV plasma cell

Host anti-CMV plasma cell

Prevention of CMV dissemination

Host anti-CMV B cell

Host anti-CMV

B cell

Donor alloreactive

CD8+ T cell CMV dissemination CMV disease

18 JANUARY 2019 • VOL 363 ISSUE 6424 233

B cells help prevent CMV reactivation

T cells, NK cells, and B cell–derived antibodies contribute to the prevention of mouse CMV dissemination

after BMT. Killing of host B cells and plasma cells by alloreactive donor T cells (GVHD) in combination with

therapeutic deletion of T cells and NK cells resulted in CMV infection of latently infected hosts.

Published by AAAS

on January 17, 2019^

http://science.sciencemag.org/

Downloaded from