Reviews: Systemic Immunity in Cancer

REVIEWS
Systemic immunity in cancer

Kamir J. Hiam-Galvez1,2,3,4,5,6, Breanna M. Allen1,2,3,4,5,6 and Matthew H. Spitzer 1,2,3,4,5,6 ✉
Abstract | Immunotherapy has revolutionized cancer treatment, but efficacy remains limited in
most clinical settings. Cancer is a systemic disease that induces many functional and compositional
changes to the immune system as a whole. Immunity is regulated by interactions of diverse cell
lineages across tissues. Therefore, an improved understanding of tumour immunology must
assess the systemic immune landscape beyond the tumour microenvironment (TME). Importantly,
the peripheral immune system is required to drive effective natural and therapeutically induced
antitumour immune responses. In fact, emerging evidence suggests that immunotherapy drives
new immune responses rather than the reinvigoration of pre-existing immune responses.
However, new immune responses in individuals burdened with tumours are compromised even
beyond the TME. Herein, we aim to comprehensively outline the current knowledge of systemic
immunity in cancer.
Peripheral immune cells

Cancer is a systemic disease, and prolonged inflammation of immune responses to cancer must encompass all
Immune cells in an individual is a hallmark of cancer1. Whether this inflammation ini- immune cell lineages across the peripheral immune
bearing a tumour that are tiates tumorigenesis or supports tumour growth is system in addition to within the TME.
not in the tumour context dependent, but ultimately the global immune Recent clinical and preclinical studies are beginning
microenvironment.
landscape beyond the tumour becomes significantly to unravel the range of systemic immune perturbations
altered during tumour progression. Over the last dec- that occur during tumour development as well as the
ade, targeting the immune system with immunotherapy crucial contribution of peripheral immune cells to an
has revolutionized cancer therapy. Modulation of the antitumour immune response. Here, we review recent
existing patient immune system through immune check- advances that set the stage for a new holistic vantage
point inhibitors (ICIs) such as anti-CTLA4, anti-PD1 point of tumour immunology to map and therapeutically
1
Department of
Otolaryngology — Head and and anti-PDL1 has led to durable remissions across a harness the entirety of an immune response to cancer.
Neck Surgery, University wide variety of different tumour types. Moreover, infu- We outline the extensive reorganization of peripheral
of California, San Francisco, sion of expanded autologous tumour-specific T cells or immune cells that coincides with malignant tumour
San Francisco, CA, USA. outgrowth as well as the systemic immunological conse-
chimeric antigen receptor T cells has proven effective
2
Department of Microbiology in patients with leukaemia. Despite these successes, quences of conventional therapies (surgery, chemother-
& Immunology, University
of California, San Francisco,
immunotherapy remains ineffective for most patients apy, radiation). We also examine the critical contribution
San Francisco, CA, USA. with cancer2,3. To date, most immunotherapies have of peripheral immune cells to driving and sustaining
3
Graduate Program in
largely been used in patients with advanced cancers, efficacious immunotherapy responses and the capacity
Biomedical Sciences, and therefore the response rate in less advanced disease of the tumour-burdened immune system to orchestrate
University of California, remains to be fully determined. Further progress towards a new immune response. Finally, we address the utility of
San Francisco, San Francisco, more broadly effective immunotherapeutic strategies peripheral immune biomarkers in aiding the diagnosis
CA, USA.
requires a deeper understanding of the immunological and prognosis of cancer and response to therapy.
4
Helen Diller Family
relationships between tumours and their hosts across
Comprehensive Cancer
Center, University of
the body. Perturbations induced by tumour burden
California, San Francisco, The tumour immunology field has focused heavily Many human cancers and mouse models of cancer
San Francisco, CA, USA. on local immune responses in the tumour microenvi- drive extensive disruption of haematopoiesis. This dis-
5
Parker Institute for Cancer ronment (TME), yet immunity is coordinated across ruption manifests most prominently in an expansion of
Immunotherapy, San tissues. For example, many myeloid cells are frequently immature neutrophils and monocytes in the periphery
Francisco, CA, USA.
replenished from haematopoietic precursors in the bone of tumour-burdened hosts, which then also traffic to
6
Chan Zuckerberg Biohub, marrow4, and critical T cell priming events typically the TME and contribute to local immunosuppres-
San Francisco, San Francisco,
CA, USA.
occur in lymphoid tissues5. The localized antitumour sion. This phenomenon has been reviewed extensively
✉e-mail: matthew.spitzer@ immune response cannot exist without continuous elsewhere8–10. In brief, haematopoietic stem and progen-
ucsf.edu communication with the periphery. Furthermore, vir- itor cells are mobilized into proliferation and differenti-
https://doi.org/10.1038/ tually every subset of immune cell has been implicated in ation towards the monocytic and granulocytic lineages,
s41568-021-00347-z cancer biology6,7. Therefore, a thorough understanding resulting in peripheral expansion and intratumoural
nATure RevIewS | CanceR volume 21 | June 2021 | 345
0123456789();:
Reviews
Immune macroenvironment accumulation of immature immunosuppressive neutro- and proliferation in many contexts, including cancer29,30.
The total immune system phils (often referred to as polymorphonuclear myeloid- The frequencies of dendritic cell subsets are decreased in
in a tumour-burdened host derived suppressor cells (PMN-MDSCs)), monocytes peripheral blood of human ovarian31, prostate31, breast32,
comprising blood and (often referred to as M-MDSCs) and macrophages11–15 lung33 and renal34 cancers as well as head and neck squa-
secondary lymphoid organs
(Fig. 1). Mouse models of breast cancer and rhabdomyo- mous cell carcinoma35 and melanoma36 when compared
such as the bone marrow,
spleen and lymph nodes. sarcoma have demonstrated that the frequency of bone with healthy control donors (Fig. 1). In patients with pan-
marrow haematopoietic stem cells, multipotent progeni- creatic or breast cancer and in mouse models of these
Co-stimulatory and tors and granulocyte monocyte progenitors progressively cancer types, a decrease in the frequency of peripheral
co-inhibitory receptors
increase with the tumour burden12,16–18. Importantly, a type 1 conventional dendritic cells (cDC1s) was driven
Immunological receptors
expressed on the surface
pan-cancer study found elevated levels of haematopoie- through tumour-derived G-CSF, which caused a down-
of lymphocytes, antigen- tic stem cells, multipotent progenitors and granulocyte regulation of IRF8 in dendritic cell precursors, reducing
presenting cells and tumour monocyte progenitors in the blood of patients with the differentiation of mature dendritic cells18 (Fig. 1).
cells that stimulate or inhibit breast, cervical, liver, oesophageal, lung, ovarian and Similarly, tumour-derived vascular endothelial growth
immune cell functions.
gastrointestinal cancers, suggesting that haematopoietic factor (VEGF) has been shown to inhibit the maturation
Lymphopenia dysregulation is common in human cancer13. A com- of dendritic cell precursors37,38. An alternative mecha-
The reduced abundance of prehensive meta-analysis of more than 40,000 patients nism for dendritic cell paucity in a mouse model of
lymphocytes (T cells and/or found that elevated neutrophil frequencies in the pancreatic cancer was shown to be mediated by serum
B cells and/or natural killer
blood, as measured by the neutrophil to lymphocyte IL-6 driving increased dendritic cell apoptosis39 (Fig. 1).
cells) in the blood.
ratio, were associated with poor prognosis in patients In patients with pancreatic cancer and mouse models of
with mesothelioma, pancreatic cancer, renal cell carci- pancreatic cancer, peripheral dendritic cells differenti-
noma, colorectal carcinoma, gastroesophageal cancer, ate into a semi-mature state characterized by moderate
non-small-cell lung cancer (NSCLC), cholangiocarcinoma increases in co-stimulatory and co-inhibitory receptors39–41
and hepatocellular carcinoma19. Several factors have been (Fig. 1). Bulk transcriptomic analyses of these peripheral
implicated in driving this process including G-CSF12,20, dendritic cells from mice bearing pancreatic tumours
GM-CSF17,21,22, IL-17 (ref.15), oxysterol23, IL-8 (ref.24), revealed that these semi-mature dendritic cells showed
CCL2 (ref.14), TNF25, tumour-derived exosomes26 and upregulation of genes involved in proteasomal degrada-
IL-1β (ref.27). tion but did not show upregulation of T cell polarizing
The vast majority of research that highlights periph cytokines39, suggesting that, similar to semi-mature den-
eral immune perturbations in the context of cancer has dritic cells in other contexts, they only partially possess
focused on this increase in immature and immuno the capacity to provide stimulation to T cells.
suppressive myeloid populations; however, this expan- Substantially less is known about the organization of
sion also often co-o ccurs with alterations to many other major immune lineages in the tumour macroenvi-
other peripheral immune lineages (Fig. 1). Our group ronment. Lymphopenia is common in patients with breast
recently used mass cytometry to comprehensively pro- cancer, lymphoma and sarcoma42. Interestingly, circulat-
file the phenotype and frequency of all major immune ing T cells in patients with breast43, lung44 and cervical45
lineages in the bone marrow, spleen, blood, draining cancers have decreased diversity in the repertoire of
lymph node (dLN) and tumour in eight distinct mouse T cell receptors (TCRs) (Fig. 1). As greater TCR diver-
tumour models28. Although we observed a periph- sity is associated with better tumour control in patients
eral myeloid expansion in all tumour models as has with melanoma46, an improved understanding of TCR
been previously described, we also observed extensive repertoire fluctuations driven by cancer is warranted.
peripheral immune reorganization across lineages and Furthermore, as a decreased TCR repertoire in humans
tissues. For example, three models of breast cancer is associated with age47 as well as other prior immunolog-
(AT3, 4T1, MMTV-PyMT) spanning three different ical exposures such as chronic infection48, these changes
mouse strain backgrounds showed extensive splenic may also be a cause for malignant outgrowth. The causal
immune population remodelling characterized by relationship between TCR diversity and cancer has yet
phenotypic shifts as well as increased frequencies of to be determined. Peripheral T cells are also function-
neutrophils, eosinophils and monocytes along with ally perturbed, as polyclonal memory CD4+ and CD8+
reductions in dendritic cell, B cell and T cell popula- T cells from peripheral blood have decreased capacity to
tions (Fig. 1). Strikingly, surgical resection of the tumour produce IL-2 and IFNγ in response to stimulation with
or cytokine blockade treatments reversed many of the PMA and ionomycin in human patients with breast
changes, suggesting plasticity in the peripheral reorgan- cancer49. Peripheral naive CD4+ T cells also exhibited
ization of the immune macroenvironment in cancer. These decreased responses to IL-6 stimulation as measured by
data demonstrate that tumour development dramati- phosphorylation of STAT1 and STAT3 in patients with
cally restructures the global immune landscape across breast cancer50 (Fig. 1).
immune cell lineages. The most studied perturbation of T cells in cancer
Beyond excessive production of monocytic and is the expansion of suppressive CD4+ regulatory T (Treg)
neutrophilic cells through aberrant haematopoiesis, per- cells in the periphery and their infiltration into the
turbations in dendritic cells have been observed in the tumour51. Recent work has now shown that Treg cells
periphery of tumour-burdened hosts. This has impor- present in the blood of patients with cancer share pheno
tant implications for the development of antitumour typic and TCR repertoires with intratumoural T cells,
immune responses, as dendritic cells are critical orches- suggesting that a significant proportion of intratumoural
trators of CD8+ and CD4+ T cell priming, differentiation suppressive Treg cells are derived from naturally occurring
346 | June 2021 | volume 21 www.nature.com/nrc
0123456789();:
Reviews
a Haematopoietic stem
and progenitor cells
VEGF Common myeloid Common lymphoid
progenitor progenitor
G-CSF
Bone ↓ NF-κB
marrow ↓ IRF8
Erythrocytes Mega- Basophil Eosinophil GMP ↑ STAT3 Common T cell B cell NK cell
karyocyte dendritic cell
progenitor
Unchanged or
impact unknown Immature Immature Plasma cell
Pre-dendritic cell
Increasing monocyte neutrophil
Decreasing
Multiple
development
steps
Monocyte Neutrophil Dendritic cell
b ↑ Frequencies ↓ Frequencies Altered function
Blood PMA and

ionomycin
CD8 + ↓ TCR
T cell clonality
IL-10 HSC MMP GMP IL-6
Cytotoxic
Clonal CD4+ granules
Treg cell T cell
↓ pSTAT1
Semi-mature ↑ Apoptosis ↓ pSTAT2 ↓ IFNγ ↓ Cytotoxicity
dendritic cell
c Spleen d Tumour dLN

↑ Frequencies ↑ Frequencies
↓ Frequencies ↓ Frequencies
Fig. 1 | Systemic perturbations to immune organization by the the periphery before infiltrating the tumour. Dendritic cell as well as
tumour burden. The peripheral immune landscape is perturbed in CD8+ and CD4+ T cell frequencies and T cell receptor (TCR) repertoire
many tumour types. The bone marrow, blood, spleen and draining lymph diversity are decreased in many tumour contexts. Functional deficits in
node (dLN) form an immunological network in constant communication response to stimuli have been identified in T cell populations. CD4 +
during tumour development. a | Bone marrow haematopoiesis skews T cells exhibit decreased signalling responses to IL-6 stimulation, and
towards the production of neutrophils and monocytes through increased both CD4+ and CD8+ T cells produce less IL-2 and IFNγ in response to
frequency of haematopoietic stem cells (HSCs) and granulocyte PMA and ionomycin stimulation. Natural killer (NK) cell cytotoxic
monocyte progenitors (GMPs). In some contexts, this skewing occurs at potential is also decreased. c | Several alterations observed in the blood
the expense of dendritic cell precursors which share progenitors, have been mirrored in the spleen in mouse models, including
leading to a systemic paucity of dendritic cells that has been shown to accumulation of immature neutrophils, monocytes and semi-mature
be driven by G-CSF stimulating STAT3 signalling while repressing IRF8, dendritic cells. Decreased abundance of dendritic cells and T cell
as well as through vascular endothelial growth factor (VEGF) decreasing populations has also recently been described. d | The tumour dLN has the
NF-κB signalling. T cell, B cell and plasma cell populations in the bone most direct line of communication with the tumour and is characterized
marrow have also been shown to be decreased. b | During tumour by increased frequency of monocytes and dendritic cells with a decrease
development, bone marrow progenitor pools as well as suppressive of CD8+ T cells. Collectively, these observations across many human and
immature monocytes and neutrophils are mobilized into circulation in mouse tumour models demonstrate that the peripheral immune
the blood. Systemic increased frequencies of suppressive lymphocyte landscape is shifted towards a suppressive state marked by increases in
populations, CD4+ regulatory T (Treg) cells and regulatory B cells are also anti-i nflammatory cell types and decreases in key mediators of
commonly observed. Treg cells specifically undergo clonal expansion in antitumour immunity. MMP, multipotent progenitor.
0123456789();:
Reviews
thymic Treg cells rather than through tumour-induced Changes induced by conventional therapy
differentiation of naive CD4+ T cells52,53 (Fig. 1). Conventional therapeutic strategies in cancer, includ-
Another suppressive lymphocyte population ing chemotherapy, radiation and surgery, perturb the
that plays a role in tumour progression is regulatory global immune landscape. Understanding these sys-
B cells, which are characterized by production of the temic immune consequences is important for designing
anti-inflammatory cytokine IL-10 (ref.54). Similar to strategies that augment rather than impede antitumour
Treg cells, an expansion of IL-10-producing regulatory immune responses, which can include optimal timing,
B cells has been documented in peripheral blood of dosing or combinations.
patients with gastric cancer55,56 and patients with lung
cancer57, whereas frequencies of total B cells remained Chemotherapy and radiation therapy remodel circulat-
unchanged (Fig. 1). In the 4T1 mouse model of breast ing immune populations. Chemotherapy and radiation
cancer, suppressive CD25 + regulatory B cells were therapy are designed to target cancer cells by compro-
expanded in the spleen, lymph nodes and blood58. mising cellular integrity during division; however, these
Natural killer (NK) cells are yet another important agents can also induce remodelling of immunity that
component of antitumour immunity that can directly can either impede or augment overall treatment effi-
kill tumour cells as well as influence antitumorigenic cacy. Consequences of conventional cancer therapies
behaviour of other immune cells 59. Peripheral NK on the immune system were well reviewed by Shaked,
cells from patients with breast cancer also have altered such as expansion of immunosuppressive myeloid cells
phenotypes, characterized by decreased expression via elevated pro-inflammatory cytokines, including
of activating receptors, including NKp30, NKG2D, IL-6, IL-8 and GM-CSF, and B cell release of systemic
DNAM-1 and CD16, and increased expression of the extracellular vesicles that impede antitumour cyto-
inhibitory receptor NKG2A, as well as impaired capacity toxic immune functions65. One counter-strategy is to
to directly kill target cells and degranulate in vitro60. In pair these therapies with agents that block immuno-
patients with gastrointestinal stromal tumours, periph- suppressive phenotypes, such as inhibiting CSF1R or
eral NK cells showed decreased expression levels of the CCR2. Chemotherapeutic cytotoxicity also leads to
activating receptor NKp30 and impaired degranulation general lymphodepletion, and although the numbers
upon NKp30 cross-linking. Paradoxically, NK cells of CD8+ T cells in peripheral blood fully rebound to
from patients with gastrointestinal stromal tumour normal frequencies within a year, an abnormal bias
produced more IFNγ upon either IL-2 stimulation of memory CD4+ T cells towards inflammatory effec-
or incubation with dendritic cells, the latter of which tors persists for years in patients with breast cancer66.
predicted improved response to imatinib mesylate Selecting agents that mitigate immune abnormalities
treatment61. In NSCLC, transcript levels of activating may be optimal for enabling the strongest antitumour
receptors NCR1, NCR2 and NCR3 in peripheral NK immune response.
cells are all decreased, reflecting impaired natural killer The impact of chemotherapy and radiotherapy on the
cell activation, and low NCR3 transcript expression and immune system depends highly on context, making it
elevated serum levels of its ligand, B7-H6, were asso- challenging but imperative to understand how each cyto-
ciated with poor survival62. In neuroblastoma, patients toxic therapy may compromise immune function across
with metastatic disease exhibited lack of expression cancer settings. In NSCLC, standard prolonged low-dose
of the A and B isoforms of NKp30 in peripheral NK radiotherapy, but not chemotherapy, led to myeloid
cells when compared with patients with localized dis- cell expansion, reduced antigen-presenting cell func-
ease and healthy individuals63. Conversely, another tion and impaired T cell responses67. Similar immune
study in patients with NSCLC found that peripheral impacts were observed after combination chemother-
NK cells showed no phenotypic alterations by flow apy and radiotherapy in patients with cervical cancer68.
cytometry compared with cells from healthy indi- Neoadjuvant chemotherapy prior to surgical resection is
viduals, but ex vivo incubation of these NK cells with a strategy often used in breast cancer, but patients show
tumour cells induced reduction of NK cell receptor disparate immune effects depending on the cancer stage
expression and impaired degranulation compared and therapeutic agent. In patients with non-metastatic
with healthy donor-derived NK cells, suggesting that, breast cancer, doxorubicin and cyclophosphamide
in some contexts, NK cell perturbations are specific chemotherapy led to elevated numbers of circulating
to the TME64. PMN-MDSCs but no changes in M-MDSCs when com-
Altogether, these data strongly support the notion pared with pretreatment numbers69. However, in patients
that systemic corruption of immune organization occurs with metastatic breast cancer treated with 5-fluorouracil
across diverse tumour types (Fig. 1; Table 1). Further (5-FU), epirubicin and cyclophosphamide (FEC) or
work is needed to fully characterize the distinct types of docetaxel chemotherapies, M-MDSCs were dramati-
immune states in patients with cancer and the associa- cally reduced in six out of ten patients when compared
tions of these types of immune states with the tumour with pretreatment levels70. Specifically, in patients with
tissue of origin, stage of development and patient demo- breast cancer with tumours expressing HER2 (HER2+),
graphics in order to inform therapeutic development a recent study suggests that higher circulating IL-10 and
and future mechanistic studies of the causes of sys- classical monocytes associate with reduced pathological
temic disruptions. It is also critical to understand why complete responses after chemotherapy71. Future stud-
systemic immune changes are quite dramatic in some ies with larger numbers of patients and more complete
contexts yet subtle in others. measurements of the immune system are needed to
0123456789();:
Reviews
Table 1 | Peripheral immune perturbations in cancer

Immune cell type Change Tumour type and speciesa Refs
Haematopoietic stem Increased frequency Human: breast cancer, cervical cancer, liver cancer, 12,13,16,17
cells oesophageal cancer, lung cancer, ovarian cancer,

gastrointestinal cancer
Mouse: breast cancerM, rhabdomyosarcomaM
Multipotent Increased frequency Human: breast cancer, cervical cancer, liver cancer, 12,13,16,17
progenitor cells oesophageal cancer, lung cancer, ovarian cancer,

gastrointestinal cancer
Granulocyte Increased frequency Human: breast cancer, cervical cancer, liver cancer, 12,13,16,17
monocyte oesophageal cancer, lung cancer, ovarian cancer,

progenitors gastrointestinal cancer
Dendritic cell Decreased frequency Human: breast cancerNM, pancreatic cancer 18
precursors Mouse: breast cancer, pancreatic cancer

Immature Increased frequency Human: breast cancer, pancreatic cancer, lung 11,12,18,
neutrophils/ cancerM/NM, bladder cancerM/NM, head and neck 28,160–163
PMN-MDSCs cancerM/NM, glioblastoma, melanoma

Mouse: breast cancerM, melanomaM, pancreatic cancerM,
colon cancerNM, glioblastoma
Immature Increased frequency Human: renal cell carcinoma, prostate cancer, prostate 14,22,28,
monocytes/M-MDSCs cancer, melanomaM/NM, hepatocellular carcinoma 164–167
Mouse: breast cancer , melanoma , pancreatic

M/NM M
cancerM, colon cancerNM, glioblastoma

Dendritic cells Decreased frequency Human: ovarian cancer, prostate cancer, breast cancer, 11,18,28,
head and neck squamous cell carcinoma, melanomaM/NM, 32–34,36,168
lung cancerNM, renal cancer

Mouse: pancreatic cancerM, breast cancerM/NM,
glioblastoma
T cells Decreased TCR repertoire Human: breast cancerM, lung cancerM/NM, cervical cancer 43–45
Treg cells Expansion Human: lung cancerM/NM, prostate cancerNM, gastric 28,51,
cancerM/NM, colorectal cancer, oesophageal cancer, 169–174
hepatocellular carcinoma, pancreatic cancer, breast

cancerNM
Mouse: melanoma, pancreatic cancer
Treg cells Clonal expansion Human: melanomaM, gastrointestinal cancer, ovarian 52,53
cancer, breast cancer

Regulatory B cells Increased frequency Human: gastric cancerM/NM, lung cancer 55–57
Mouse: breast cancer M
CD4+ and CD8+ Decreased IL-2 and IFNγ Human: breast cancerM/NM 49
T cells production after PMA and

ionomycin stimulation
CD4+ T cells Decreased pSTAT1 and Human: breast cancer 50
pSTAT3 signalling after

IL-6 stimulation
Natural killer cells Decreased activating Human: breast cancerM/NM, lung cancerNM, 60–64
receptors, increased gastrointestinal cancerM/NM, neuroblastomaM

inhibitory receptors,
decreased cytotoxic
potential
M-MDSC mononuclear myeloid-derived suppressor cell; PMN-MDSC, polymorphonuclear myeloid-derived suppressor cell;
TCR, T cell receptor; Treg cell, regulatory T cell. aSuperscript M indicates that this observation was specifically made in metastatic
disease, whereas NM indicates that the observation was specifically made in non-metastatic disease.
parse how the disease type and stage affect the immune neoadjuvant chemotherapy in triple-negative breast
consequences of cytotoxic therapies. cancer (TNBC) induces the recruitment of new T cell
When demonstrably effective, chemotherapy can clones to the TME rather than expanding those already
augment systemic antitumour immunity in con- present72. Importantly, different subtypes of breast cancer
junction with disrupting cancer cell division. Recent showed differential immune responses to this therapeu-
work showed that effective responses to pre-surgical tic strategy, reflected in the functionality of peripheral
0123456789();:
Reviews
CD8+ T cells. Patients with oestrogen receptor-positive stress, in which mice underwent an abdominal laparo
(ER+) breast tumours had a drop or stasis in the poly- tomy and left nephrectomy, showed that surgery led to
functionality of circulating PD1+CD8+ T cells, meas- decreased systemic NK cell frequencies and tumour
ured by cytokine production after TCR stimulation. killing potential, which culminated in impaired con-
Patients with ER+HER2+ breast tumours showed a trol of lung metastasis82. Patients with colorectal cancer
complete loss of functionality in this subset. Conversely, also exhibited decreased IFNγ secretion from periph
patients with TNBC showed elevated PD1+CD8+ T cells eral NK cells when compared with healthy individ
with high functionality, producing effector cytokines uals, and this was further decreased for up to 2 months
including IFNγ and TNF, and the cytolytic molecule following surgery83. Collectively, these data suggest
granzyme B, and with evidence of clonal expansion. multiple mechanisms by which surgical resection
Ultimately, tumour-infiltrating T cells were only prog- induces global immunological perturbations that can
nostic for overall survival in TNBC. Moreover, a cytoly promote metastasis.
tic but exhausted CD8+ T cell signature in the blood However, our group recently demonstrated that the
of patients with TNBC following chemotherapy was primary tumour can be the main driver of systemic
associated with ongoing disease and was predictive of immune remodelling: successful primary tumour resec-
recurrence or metastasis post surgery. tion in mouse models of breast and colon cancer was
With the advent of immunotherapy, therapeutic sufficient to largely restore normal systemic immune
strategies are shifting towards utilizing cytotoxic ther- organization to immune cell frequencies comparable
apeutic agents that can augment antitumour immunity, with healthy control mice across the spleen, lymph node,
such as by disrupting the tumour stroma or by releasing blood and bone marrow given that there was sufficient
tumour antigens for de novo activation of the adaptive time to recover from postoperative complications28. In
immune system73–75. the highly metastatic 4T1 mouse model, lung meta
static outgrowth was observed in some animals fol-
Tumour resection can impact immunological control of lowing surgery and yet only minor systemic immune
cancer. Recent studies have provided a deeper under- changes were maintained. Furthermore, we and others
standing of the impact of surgical tumour resection on found that surgical tumour resection in mice ultimately
the systemic immune state and immunological control restored functional orthogonal responses to infec-
of metastases. Metastatic outgrowth following surgical tion, vaccination or allogeneic tumour challenges28,84.
tumour resection has been documented in several cancer Therefore, it is likely that surgery results in both detri-
types, where a wide range of pro-tumorigenic processes, mental and beneficial effects on the systemic immune
including shedding of tumour cells into circulation and system. Immunosuppressive mechanisms coinciding
stimulation of angiogenesis, lead to new and acceler- with wound healing early after surgery potentially pro-
ated metastatic growth despite resection of the primary vide a window of opportunity for disseminated cancer
tumour76. For the purposes of this Review, we focus on cells to grow out. However, the reduced primary tumour
alterations to antitumour immunity following surgery burden can ultimately restore systemic immune capac-
(other aspects driving postoperative metastasis are ity for strong adaptive responses. It will be important to
reviewed elsewhere77). Several recent studies implicate discover how the cancer type and, particularly, disease
myeloid immune cell remodelling induced by systemic stage influence immune remodelling following surgery
wound healing programmes. Resection, or wounding and the resulting potential for metastasis.
independent of primary tumour removal, triggers heal- Appropriate pairing of conventional therapies with
ing programmes that elevate circulating IL-6, G-CSF immune modulation can be a powerful tool to combat
and CCL2, and ultimately drive myeloid subsets towards cancer, and taking the systemic immune context into
immunosuppressive states78. Although resection sub- account is likely to result in improved outcomes. It is
stantially reduces the number of systemic MDSCs in the particularly important to consider the immunologically
spleen, blood and lung in the 4T1 breast cancer model, vulnerable period of time following surgery and further
functional immunosuppressive PMN-MDSCs can per- investigate the mechanisms driving these states as well
sist in these peripheral tissues for 2 weeks79. Persistent as potential therapeutic interventions to restore immune
immunosuppressive myeloid cells were shown to sup- function and prevent tumour recurrence and metastasis.
port pro-tumorigenic niches in the lungs in both breast
cancer and osteosarcoma models79,80. Mechanistically, Systemic responses in immunotherapy
one study showed that neutrophil extracellular traps Cancer immunotherapy has radically expanded our
Neutrophil extracellular increased in the liver following surgical intervention toolkit against cancer, with current US Food and Drug
traps and ensnared tumour cells to promote metastasis77. Administration (FDA) approval of 7 ICIs across 19
Extracellular web-like Modulation of myeloid subsets in the adjuvant setting different cancer types, in addition to chimeric anti-
structures comprising DNA and
can prevent post-surgical metastases, including the use gen receptor T cells, bispecific T cell engager (BiTE)
cytosolic and granule proteins
created by neutrophils to trap of gemcitabine to deplete PMN-MDSCs79 or gefitinib therapies and vaccines. The prevailing view of cancer
and neutralize invading to alter inflammatory macrophage states by blocking immunotherapy efficacy has centred around the notion
pathogens. signalling through receptor-interacting protein kinase 2 of reinvigorating cytotoxic effectors within the TME, but
(RIPK2)80. Surgical procedures in a mouse tumour appreciation is growing in the field for the fundamen-
Orthogonal responses
Immunological challenges that
vaccination model showed that tumour-specific T cell tally systemic nature of effective antitumour immunity.
do not share any antigens with responses were dramatically weakened for 7–10 days fol- Recent studies demonstrate that ICIs, including block-
the tumour. lowing surgery81. Similarly, a mouse model of surgical ade of the PD1 and PDL1 axis, rely on systemic immune
0123456789();:
Reviews
Box 1 | Microbiome modulation of systemic immunity in cancer blood) after productive antitumour responses is suffi-
cient to protect naive animals92. This same study showed
The activity and composition of the microbiome influences the organization of the that systemic PDL1 blockade can break tolerance to dis-
human immune system175. Antibiotic treatment that disrupts the gut microbiome seminated tumours when paired with local therapeutic
leads to resistance to immune checkpoint inhibitors (ICIs) in mouse models of cancer delivery at one site.
and in patients with cancer176. Multiple studies have found that faecal microbiota
It has become clear that inhibiting the PD1–PDL1
transplantation (FMT) from patients into mouse models can recapitulate functional
outcomes on tumour control and response to ICIs such that FMT from ICI responders axis has impacts beyond blocking local immunosuppres-
drives improved antitumour immune responses compared with FMT from ICI non- sive cues in the tumour, and recent work has clarified
responders176–180. In fact, FMT from ICI non-responders compared with responders drove key peripheral immune cells driving responses in these
divergent peripheral immune responses, such as higher frequencies of regulatory CD4 settings. First, therapeutic benefit of immune check-
T cells and T helper 17 cells in the spleen of non-responder FMT recipients, suggesting point inhibition is only observed in models with intact
systemic consequences of microbiome composition in patients with cancer177. Although host PD1 and PDL1 expression and is less dependent
much mechanistic work is still needed to link microbiome to immune composition and on cancer cell expression of PDL1 (refs94–96). Aside from
function, Bifidobacterium pseudolongum and Akkermansia muciniphila have been shown tumour cells, the majority of cells that express PDL1 are
to produce inosine that activates antitumour T cells via the adenoside A2A receptor179. antigen-presenting cells, including macrophages and, at
One study identified an 11-strain mixture of commensal bacteria that enhances
even higher levels, cDCs97. In patients with melanoma
antitumour immune responses through CD103+ dendritic cell-orchestrated CD8+ T cell
responses. Systemically, the same 11-strain mixture also drove enhanced intestinal or ovarian cancer, expression levels of PDL1 on intra-
bacterial clearance following oral Listeria monocytogenes infection as well as improved tumoural macrophages and cDCs correlate with clini-
spleen and liver bacterial clearance after intraperitoneal L. monocytogenes infection180. cal complete responses to anti-PDL1 and anti-CTLA4
These data suggest that microbiome-based improvements of antitumour immunity also therapy 95. Moreover, several groups have recently
shape systemic immunity. This topic has been reviewed in greater depth elsewhere181. demonstrated that dendritic cells are a critical media-
tor of PDL1 blockade efficacy97–100. Targeted depletion
mechanisms to achieve effective antitumour responses. of PDL1 in cDCs, but not macrophages, substantially
Furthermore, the microbiome is emerging as a potent reduced CD8+ T cell responses and tumour shrinkage
modulator of the immune system, and the relevant in response to PDL1 blockade in the subcutaneous
impact on antitumour immune responses is discussed MC38 mouse cancer model97. The critical location
in Box 1. for this interaction appears to be the tumour dLNs, as
tumour-specific PD1+ T cells in the dLN showed high
Intact peripheral immunity is critical for immunother- co-localization with PDL1-expressing cDCs98. Selective
apeutic efficacy. Intact peripheral immune function, targeting of PDL1 engagement in the dLNs was suffi-
communication and trafficking are required for ICI cient to induce effective antitumour responses across
efficacy. Disruption of peripheral immune integrity by two syngeneic models, albeit to a lesser extent than sys-
systemic chemotherapy can impede therapeutic benefit temic PDL1 blockade98. Further supporting the signifi-
by PD1 blockade, causing systemic lymphodepletion cance of PDL1 activity specifically on cDCs, interactions
and abrogating long-term immune memory85. By con- between PDL1+ cDCs and PD1+ T cells in dLNs were
trast, local chemotherapy spares peripheral immunity, indicative of the disease dissemination status in patients
collaborating with PD1 blockade to induce dendritic with melanoma. Frequent PD1 and PDL1 interactions
cell infiltration into the tumour and clonal expansion in dLNs were observed in patients with metastatic mel-
of antigen-specific effector T cells85. A specialized sub- anoma and were predictive of early disseminated disease
set of CD103+ dendritic cells transport tumour antigen recurrence in patients with non-metastatic melanoma98.
to the peripheral immune system by CCR7-dependent Augmenting effector and memory T cell development in
migration from the tumour to the dLN, where the the dLN via mitochondrial activation further improved
priming of tumour-specific CD8+ and CD4+ T cells PD1 blockade efficacy in tumour-bearing mice101, again
occurs86–89 (Fig. 2). cDC2s are also capable of trafficking highlighting that systemically engaged immunity is
tumour antigen to the dLN and priming tumour-specific clearly optimal for tumour eradication.
CD8+ and CD4+ T cells; however, this process is often
restrained by intratumoural Treg cells30,90. Recent evidence Effective immunotherapies drive de novo immune
suggested that dendritic cells migrating from the tumour responses. Productive antitumour responses ultimately
to the dLN can transfer antigen to lymph node-resident necessitate functional effector lymphocytes within the
dendritic cells that can then also prime tumour-specific TME to mediate cancer cell killing. However, recent
T cells90 (Fig. 2). Newly primed tumour-specific T cells studies revealed that intratumoural T cells acquire
then traffic from the lymph node to the tumour, and terminally exhausted states over time, rendering them
this cycle is an essential process in natural and thera- incapable of key effector functions (Fig. 2). Analysis
peutically induced antitumour immunity91. As further of epigenetic landscapes of CD8+ T cells in mice and
evidence of the systemic nature of antitumour immunity, patients with cancer showed that intratumoural T cells
blockade of lymphocyte egress from lymphoid organs or underwent extensive chromatin remodelling, which
surgical resection of tumour dLNs abrogates immuno locked cells in dysfunctional states and reduced the
therapeutic efficacy92,93. The eradication of systemic ability of these cells to produce TNF and IFNγ102. This
disease also heavily relies on global immune responses. process was biphasic in preclinical models, where early
Strong adaptive immune responses confer peripheral T cell remodelling was reversible upon removal from the
memory, where the transfer of T cells from secondary tumour context, but a second wave of epigenetic remod-
lymphoid organs (including the spleen, lymph node and elling led to irrecoverable T cell dysfunction marked by
0123456789();:
Reviews
a Baseline treatment-naive tumour dLN

b After anti-PDL1 or anti-CD40 immunotherapy
Antigen
transfer
Blood
Proliferating Proliferating
T cells T cells
cDC with
tumour antigen
Infiltrating
cDC with
no tumour
antigen
Tumour
PD1 CD86 CD40

New active Dysfunctional Lymph node
Macrophage Anti-CD40
CD8+ T cell CD8+ T cell resident
dendritic cell
New active Dysfunctional

Migratory Tumour antigen Anti-PDL1
CD4+ T cell CD4+ T cell
dendritic cell PDL1 CD28 CD40L
Fig. 2 | Systemic immune responses in cancer immunotherapy. Effective therapy must rely on another source of functional effector T cells.
responses to immunotherapy drive de novo peripheral immune responses. b | Immunotherapeutic intervention through PD1 and PDL1 checkpoint
Schematic illustrating how functional antitumour responses are reliant on blockade increases the interaction between cDCs and naive T cells in the
immune dynamics outside the tumour microenvironment (TME). a | At dLN, and, alongside CD28 co-stimulation, facilitates the priming and rapid
baseline, conventional dendritic cells (cDCs) in the TME take up tumour expansion of new T cell clones with new antigen specificities. Checkpoint
antigen and travel to the draining lymph node (dLN), where they can then blockade also leads to the proliferation of existing T cell clones in
transfer antigen to resident cDCs through the formation of direct synapses. circulation. These expanding peripheral T cells ultimately infiltrate the TME,
T cells in the TME reach states of terminal exhaustion due to chronic and express markers indicative of antigen-s pecific activation and
stimulation, the harsh environment and immunosuppressive cues. demonstrate functional cytotoxicity. Productive de novo immune responses
Dysfunctional intratumoural T cells accumulate structurally damaged can also be achieved through CD40 agonism, which can drive cDC
mitochondria, and upregulate CD103 and CD38 coinciding with irreversible activation in settings resistant to checkpoint blockade and initiate these
epigenetic remodelling. Thus, effective antitumour responses driven by new T cell responses to replace exhausted intratumoural clones.
CD101 and CD38 co-expression102. Additional studies and cell survival transcriptional and epigenetic pro-
have identified the transcription factor TOX as a critical grammes compared with exhausted CD8+ T cells107.
regulator in transcriptional and epigenetic reprogram- These productive intratumoural CD8+ T cells can be
ming in response to chronic T cell stimulation, leading identified by expression of the transcription factor TCF1
to T cell exhaustion103,104. This intratumoural T cell dys- (which is involved in WNT signalling in stem cell-like
function is driven by microenvironmental stressors in memory programmes) and notable lack of expression of
the TME, chronic TCR stimulation and checkpoint CD39 and TIM3 (ref.107). Thus, immunotherapy efficacy
protein signalling103–106. A study recently showed that relies on the quality of effector CD8+ T cells within the
metabolic challenges within the tumour can cause TME, but persistence in this toxic microenvironment
T cells to accumulate structurally damaged mitochon- rapidly drives dysfunctional differentiation of T cells
dria with high levels of reactive oxygen species and that lose their ability to efficiently contribute to tumour
T cell exhaustion
overall compromised membrane potential105. Importantly, clearance.
A terminal state of T cell
differentiation driven by mitochondrial dysregulation was sufficient to induce To overcome local immune dysfunction, effective
chronic T cell receptor epigenetic reprogramming towards terminal dysfunc- immunotherapies drive de novo peripheral immune
(TCR) stimulation and tion and was not observed in peripheral T cells from responses culminating in new effector T cell infiltra-
characterized by expression the spleen or dLN. Clinical responses to ICIs in patients tion (Fig. 2). Several reports have now shown that PD1
of inhibitory receptors and
hypofunctionality, including
with melanoma were associated with the presence of a and PDL1 blockade drive novel T cell clones into the
a reduced capacity to stem-like CD8+ T cell state with reduced expression of TME that were not present locally prior to therapy108–110.
secrete cytokines. co-inhibitory molecules and elevated memory, activation In a recent study in patients with basal cell carcinoma
0123456789();:
Reviews
before and after PD1 blockade, 68% of all intratu- shows a greater than 50% objective response rate and
moural CD8+ T cell clones after PD1 blockade were the induction of proliferating CD4+ and CD8+ T cells
novel, 84% of which displayed exhaustion markers in the blood119. CD40 activation is just one strategy
indicative of antigen-specific activation, and these demonstrating that converting immunologically ‘cold’
cells represented novel TCR specificity groups, sug- tumour contexts into ‘hot’ immune involvement requires
gesting priming against new antigen targets108. Further, de novo immune responses rather than reinvigoration.
35.5% of these novel clones were also detected in the
blood, with 11.8% detected in circulation pretreat- Secondary immune challenges in cancer
ment. Correlation between T cell clones in the blood The prior experience and state of the immune system
and tumour was also demonstrated in patients with dramatically shapes future responses to new challenges.
metastatic melanoma, renal cancer, lung cancer and Altered basal cytokine levels, cellular composition
colon cancer109,110. Anti-CTLA4 has also been shown and cellular activation states are known to impact the
to dramatically increase peripheral T cell reactivities in nature and magnitude of secondary responses in mod-
patients with melanoma, suggesting new T cell priming els of chronic infection and co-infection120–122. As the
as a mechanism of action111. systemic immune state is significantly reorganized in
Mechanistically, cell-intrinsic CD28 signalling in individuals bearing tumours, this may have functional
CD8+ T cells is critical in PD1 blockade efficacy112,113, consequences on the orchestration of new immune
providing necessary co-stimulation for naive T cell responses. Identifying systemic functional deficits to
priming. In line with this finding, a higher baseline immunological challenges, such as vaccines or infec-
proportion of CD28–CD57+KLRG1+ senescent CD8+ tions, in patients with cancer remains challenging due
T cells in the blood of patients with NSCLC was associ- to the effects of common cancer therapies. Patients with
ated with resistance to ICIs114. Impressively, in patients cancer are capable of developing detectable antibodies
with classical Hodgkin lymphoma, the peripheral T cell in response to influenza vaccination that are comparable
clonal diversity at baseline was associated with PD1 with healthy individuals123. However, during the ongo-
blockade efficacy, illustrating how individual systemic ing 2020 SARS-CoV-2 pandemic, patients with cancer
immune contexts dictate the impact of immunothera- who are infected with SARS-CoV-2 are more likely to
peutic intervention115. This signature was complemented develop severe symptoms and exhibit higher mortality
by greater expansion of singleton clones in the blood rates124,125. Notably, even infected patients not receiving
of patients with complete response, likely representing cancer treatment were at increased risk of mortality and
peripheral T cells that had not encountered antigen pre- severe illness125. This observation suggests that the sub-
treatment. The antitumour immune response in this stantial phenotypic and compositional changes to the
cancer context was more reliant on CD4+ T cells, with systemic immune system across many cell types could
expanded CD4+ TCR diversity, and concordant associa- lead to altered immune responses to a secondary chal-
tions with circulating B cell abundance and a novel innate lenge outside the TME. As an intact functional periph-
effector population capable of antibody-dependent cel- eral immune system is critical for the development of
lular cytotoxicity. Together, these results support the new antitumour immune responses, as described in the
notion that not only is peripheral immunity involved in previous section, it is imperative to understand how
renewed antitumour responses but also de novo priming immunological decisions are made within the context
of additional naive T cells with new antigen specificities of a tumour-burdened state.
contributes to effective immunotherapy (Fig. 2). Recent work has begun to mechanistically dissect
Intentional strategies for driving de novo immune why the tumour-burdened immune state results in
responses are gaining traction in clinical trials, includ- weakened peripheral secondary immune responses by
ing stimulation of dendritic cell activity through vari- investigating the effects of immune challenges that share
ous strategies such as agonistic CD40 antibodies116,117. no antigens with the initial tumour. By utilizing preclini-
Several studies have demonstrated that immune check- cal mouse model systems, the many confounding factors
point blockade relies on derestricting cDCs to allow that impact studies in patients are avoided. It has been
for effective T cell priming97–100, but this strategy fails reported that of breast tumour-bearing mice mounted
in cases where there is a poor or absent pre-existing weaker antibody responses and T cell proliferation in
activation of antigen-presenting cells. Patients with response to an immunization challenge as well as showed
pancreatic cancer are resistant to ICIs, but preclini- impaired rejection of an allogeneic tumour84 (Fig. 3a).
cal models demonstrate that combination with CD40 Similarly, our group showed that mice with AT3 breast
agonism can produce complete pancreatic tumour tumours infected with Listeria monocytogenes mounted
regression and extend survival independent of TLR, diminished splenic antibacterial responses marked by
STING or IFNAR signalling118. Efficacy in this particu- decreased dendritic cell expression of CD86, CD80 and
lar KrasLSL-G12D/+,Trp53LSL-R172H/+,Pdx1–Cre model was CD83 at 2 days post infection28 (Fig. 3b). This ultimately
dependent on host BATF3+ cDC1s and CD40, as well led to reduced CD8+ T cell proliferation and differen-
as effector CD4+ and CD8+ T cells, which massively tiation at 7 days post infection compared with healthy
expanded in the blood and dLNs, indicative of peripheral control mice infected with L. monocytogenes, which
immune activation118. An early clinical trial in patients could be rescued by CD40 agonist treatment or surgical
with metastatic pancreatic cancer holds promise, where resection of the tumour (Fig. 3b) . Strikingly, surgi-
combination of CD40 agonism with PD1 blockade cal tumour resection also restored humoral and cellular
and gemcitabine and nab-p aclitaxel chemotherapy responses to immunization84. Similarly, the splenic CD8+
0123456789();:
Reviews
a Healthy control Tumour-bearing mouse T cell response to L. monocytogenes in Pan02 pancre-

atic tumour-bearing mice also showed perturbed T cell
differentiation characterized by the acquisition of an
versus exhausted fate126. Suppressed splenic expansion of CD8+
T cells has also been observed in response to lympho-
cytic choriomeningitis virus in mice with pre-existing
B16 melanoma127 or AT3 breast tumours28 when com-
pared with healthy control infected mice. Vaccination of
pancreatic tumour-bearing mice with ovalbumin (OVA)
Immunization Immunization Bacterial infection Viral Matrigel and CpG also led to impaired OVA-specific CD8+ T cell
(hen egg (ovalbumin (Listeria infection plug proliferation and differentiation in the spleen when
lysozyme) and CpG) monocytogenes) (LCMV) (ovallbumin)
compared with healthy control vaccinated mice39. This
impairment was linked to dendritic cell dysfunction
and could be rescued by combined treatment with
FLT3L and CD40 agonism to increase both dendritic
cell numbers and activation, respectively39 (Fig. 3b).
In a PyMT-B6 mouse model of breast cancer, a Matrigel
↓ Antibody response ↓ CD8+ T cell proliferation and
in tumour-bearing mice differentiation in tumour-bearing mice plug containing poly I:C and OVA was used as an immu-
nogenic secondary challenge without shared antigens
b Systemic cDC function to the primary tumour. Pre-existing malignancy drove
Healthy control Tumour bearing Tumour bearing Tumour bearing significantly decreased frequency of cDC1s within the
and anti-CD40 and FLT3L plug and the dLN, which then led to a reduced number
Dendritic cell frequency
of OVA-specific CD8+ T cells infiltrating the plug when

compared with healthy control challenged mice18. Taken
together, the results of these studies show that the innate
and adaptive arms of immune responses, and specifi-
cally dendritic cell and CD8+ T cell interactions, do not
proceed optimally in the context of cancer (Fig. 3). Thus,
Anti-
CD40 FLT3L therapeutic strategies aiming to stimulate new CD8+
T cell responses must overcome these obstacles.
CD86
Dendritic cell
CD80 Systemic immune biomarkers for cancer

activation
Despite significant interest in the development of predic-

tive biomarkers leveraging the systemic immune system,
the vast majority of immunotherapy clinical trials are
still performed without the use of a biomarker to guide
inclusion128. Currently, there is no systemic immune bio-
marker that is sufficiently established to permit bedside
decision-making, although some immunological fea-
tures in the TME have been shown to be associated with
prognosis in various contexts. Therefore, an opportunity
exists for immune biomarkers from peripheral blood to
Proliferating
CD8+ T cells
effector
help guide patient treatment decisions.
Circulating protein biomarkers. Quantification of cir-

culating proteins in the serum or plasma is routinely
performed in various pathological contexts, and thus
several studies have examined the potential of this
Fig. 3 | Secondary immune challenges in the context of cancer. Orthogonal challenges approach to develop predictive biomarkers for cancer
that do not share antigens with the tumour and drive immune responses away from the therapy (Table 2). In general, higher levels of soluble fac-
tumour microenvironment have revealed functional impairments in tumour-burdened tors associated with ongoing immune responses appear
hosts. a | Various challenges in tumour-bearing mice, including immunizations, bacterial to indicate improved prognosis. Increased levels of IL-2
and viral infections, and a Matrigel plug containing a new antigen not expressed in the and decreased levels of IL-6 and TNF in the blood at
tumour have shown blunted CD8+ T cell proliferation and differentiation. An impaired baseline as well as an increase in IL-4 levels on treatment
antibody response has also been observed in response to immunization. b | Mechanistically, were all associated with improved response to ICIs in
several of these challenges (immunization, bacterial infection and Matrigel plug) have patients with small cell lung cancer129. Moreover, a study
been linked to systemic dendritic cell paucity or impaired activation, described in the
of patients with NSCLC found that increased serum lev-
draining lymph node (dLN) and spleen. The precise drivers of dendritic cell impairment
in cancer are still being investigated, but they involve altered dendritic cell development els of numerous inflammatory cytokines were associated
and apoptosis induced by increased circulating IL-6. Immunotherapeutic interventions with improved response to anti-PD1 therapy and overall
that activate dendritic cells (anti-CD40) or increase their abundance (FLT3L) have survival130. By contrast, high serum levels of the neu-
restored CD8+ T cell proliferation and differentiation. cDC, conventional dendritic cell; trophil chemokine IL-8 have recently been associated
LCMV, lymphocytic choriomeningitis virus. with poor response to ICIs in patients with melanoma,
0123456789();:
Reviews
Table 2 | Peripheral immune biomarkers in cancer investigation. Beyond secreted factors, studies have also
examined whether the repertoire of lymphocyte anti-
Peripheral blood Prognosis Tumour typea Refs gen receptors in circulation provides prognostic infor-
immune feature
mation. Baseline diversity of TCR repertoires was also
High neutrophil to Worse response to ICIs MelanomaM/NM, 132,133,
shown to predict responses to anti-PD1 and anti-CTLA4
lymphocyte ratio NSCLC, renal cell 136–138
carcinomaM
ICIs in metastatic melanoma and classical Hodgkin
lymphoma115,134.
High serum IL-8 Worse response to ICIs Melanoma, NSCLC, 129,131
SCLCM/NM
Peripheral cellular biomarkers before treatment.
Low serum LDH Better response to ICIs MelanomaM/NM, 132,133
Cellular biomarkers from peripheral blood are another
NSCLCM
promising approach to improve patient stratification.
Higher relative eosinophil Better response to ICIs MelanomaM/NM 132
Simple metrics quantified from routine complete
count blood counts have been shown to associate with patient
Higher relative Better response to ICIs MelanomaM/NM 132
prognosis in various human malignancies. Of these,
lymphocyte count the neutrophil to lymphocyte ratio has emerged as a
Higher IL-2 at baseline Better response to ICIs SCLCM/NM 129
negative prognostic indicator in patients with various
Lower IL-6 at baseline Better response to ICIs SCLCM/NM 129 individual cancer types as well as in meta analyses19,135.
Moreover, in patients with melanoma, NSCLC and renal
Lower TNF at baseline Better response to ICIs SCLC M/NM 129
cell carcinoma, response to immunotherapy with ICIs
Increase IL-4 on treatment Better response to ICIs SCLCM/NM 129
is also associated with the neutrophil to lymphocyte
Higher TCR repertoire Better response to ICIs Melanoma , classical
M 115,134 ratio132,133,136–138. ICI response in patients with mela-
diversity at baseline Hodgkin lymphoma noma was also associated with higher relative eosinophil
Fewer M-MDSCs at Prolonged overall MelanomaM/NM 139–144 counts and higher relative lymphocyte counts compared
baseline survival after ICI (anti-CTLA4), diffuse with non-responders132. Detailed cellular analyses, gen-
(anti-CTLA4) therapy, midline glioma erally performed by flow cytometry, have also identi-
better response to (peptide vaccine +
neoantigen vaccine poly-ICLC)
fied specific cell populations in peripheral blood that
immunotherapy associate with outcomes. Various studies have identi-
fied associations between circulating immunosuppres-
More CD45RA–FOXP3hi Relapse after surgery Breast cancer 53
Treg cells at baseline sive cell subsets and response to therapy. For instance,
patients with melanoma with lower levels of circulating
More naive and effector Poor response to NSCLCM 145
Treg cells at baseline chemotherapy M-MDSCs at baseline were significantly more likely
(platinum based with or to achieve prolonged overall survival following anti-
without anti-VEGFA) CTLA4 ICI139–143. Similarly, in patients with diffuse
More Treg cells at baseline Improved response to MelanomaM 140 midline glioma, lower frequencies of M-MDSCs in
ICIs (anti-CTLA4) peripheral blood predict improved response to neo-
More mature natural killer Better response to ICIs Classical Hodgkin 115 antigen vaccine immunotherapy144. Beyond myeloid
cells at baseline (anti-PD1) lymphoma suppressive cells, a recent study demonstrated that the
baseline abundance of a subset of circulating Treg cells
More CD3–CD68+CD4+ Better response to ICIs Classical Hodgkin 115
granzyme B+ at baseline (anti-PD1) lymphoma defined as CD45RA−FOXP3hi were predictive of relapse

after patients with breast cancer underwent surgery53.
More CD127loPD1loCD4+ Better response to MelanomaM 92
T cells after treatment ICIs (anti-CTLA4 and High levels of circulating naive and effector Treg cells in
GM-CSF) patients with NSCLC were also associated with poor
response to chemotherapy145. Conversely, in the context
Proliferating and/or clonal Better response to ICIs MelanomaM 110,146,147
expansion CD8+ T cells (anti-PD1 or anti-PD1 + of immunotherapy, one study found that higher lev-
after treatment anti-CTLA4) els of circulating Treg cells predicted improved response
More Treg cells after Better response to ICIs NSCLC 148,149 to anti-CTLA4 ICI in melanoma140. On the other hand,
treatment (anti-PD1) cytotoxic cell subsets in the periphery have been shown
to associated with improved response. The abundances
Fewer PMN-MDSCs after Better response to ICIs NSCLC 148,149
treatment (anti-PD1) of mature NK cells and a subset of cells defined by

ICI, immune checkpoint inhibitor; M-MDSC, mononuclear myeloid-derived suppressor cell;
co-expression of CD68, CD4 and granzyme B were
NSCLC, non-small-cell lung cancer; PMN-MDSC, polymorphonuclear myeloid-derived found to associated with response to anti-PD1 therapy
suppressor cell; SCLC, small-cell lung cancer; TCR, T cell receptor; Treg cell, regulatory T cell; in classical Hodgkin lymphoma patients115. Therefore,
VEGFA, vascular endothelial growth factor A. aSuperscript M indicates that this observation
was specifically made in metastatic disease, whereas NM indicates that the observation was although there remains no systemic immune biomarker
specifically made in non-metastatic disease. that is widely used to guide patient treatment, recent
progress in this area has been quite encouraging.
NSCLC and small cell lung cancer129,131. This observa-
tion perhaps reflects the immunosuppressive roles that Peripheral cellular biomarkers on treatment. In addi-
neutrophils can play in antitumour immune response. tion to baseline indicators of patient outcome, various
Ovalbumin In addition, low serum LDH was found to be an indi- recent reports have emerged indicating that several
(OVA). A model antigen derived
from chicken egg whites used
cator of response to ICIs in patients with melanoma132 features of immune cells in peripheral blood of patients
to study antigen-specific T cell and NSCLC133, and the cellular and molecular mech- with cancer early after immunotherapy are indicative of
and B cell responses in mice. anisms that underlie this association merit further good outcomes. A pattern has emerged indicating that
0123456789();:
Reviews
early signs of activated or proliferating lymphocytes are immune response is regulated and dysregulated during
associated with an improved likelihood of response. effective or ineffective immune responses. Although
A study from our group identified circulating CD4+ many alterations to immune organization have been
T cells with low expression of both CD127 and PD1 observed in the periphery of individuals burdened with
associated with response to anti-CTLA4 and GM-CSF tumours, the mechanisms driving many of these features
in patients with melanoma92. In addition, various recent remain unknown. Thus, future studies will also need
studies have also shown that CD8+ T cell proliferation to provide mechanistic insights into how peripheral
and expansion in peripheral blood is associated with immune reorganization is driven in order to enable the
response to ICIs. In melanoma, the ratio of peripheral design of therapeutic strategies that restore a disrupted
T cell proliferation to tumour burden was shown to be immune system to a healthy homeostatic immune set
associated with response to anti-PD1 therapy146. In this point. Our group found that surgical resection or block-
study, the peak of CD8+ T cell proliferation occurred ade of specific cytokines in multiple tumour models
after one or two cycles of therapy (3 or 6 weeks), and restored many peripheral immune perturbations, sug-
proliferating cells (identified as Ki67+) were enriched for gesting that the tumour immune macroenvironment
PD1 expression. Building on this finding, several groups is remarkably plastic28. Pairing single-cell measure-
recently demonstrated that the expansion of specific ments from the tumour and periphery may facilitate
T cell clones in peripheral blood of patients with can- the identification of simplified biomarkers that can
cer early after ICI therapy were associated with clinical be easily sampled through blood draws and provide
responses. Clonal expansion of effector memory-like important clinical information to help guide treatment
CD8+ T cells in peripheral blood followed by tumour decisions53,115,134,156.
infiltration was associated with response to ICIs in Beyond the reorganization of the immune system in
patients with melanoma110,147. Moreover, patients across cancer, accumulating evidence also indicates that the
several malignancies with evidence of clonal expan- tumour-burdened immune state does not function in
sion by gene expression analysis experienced greater the same way as an unperturbed immune system. The
progression-free survival when treated with anti-PDL1 development of de novo antitumour immune responses
therapy109. A study of NSCLC has also found that an orchestrated from the periphery are critical for immuno
increase in circulating Treg cells after treatment with therapeutic efficacy. Therefore, any functional abnor-
anti-PD1 leads to a favourable response to treatment, malities within a tumour-burdened immune system
whereas circulating PMN-MDSC frequencies decreased may lead to suboptimal immunotherapeutic efficacy.
in responders148,149. Thus, a series of recent studies have We propose that an important avenue of future research
identified on-treatment biomarkers captured in periph- is the identification of emergent functional properties
eral blood analyses that both support the importance of of the tumour-burdened immune state. A growing
systemic immune responses in immunotherapy and pro- body of evidence suggests that systemic dendritic cell
vide opportunities for improving patient care through dysfunction is a cause of blunted CD8+ T cell prolif-
immune monitoring. eration and differentiation in the context of cancer.
Although dendritic cell-focused treatments such as
Conclusion and future perspectives anti-CD40 agonist immunotherapy28,39,157, poly-ICLC158
The widespread adoption of high-t hroughput, and FLT3L39,159 can overcome impaired dendritic cell
high-dimensional, single-cell technologies has led function and paucity, the precise mechanisms underly-
to many important discoveries and atlases of diverse ing the dendritic cell dysfunction remain incompletely
tumour immune microenvironments at steady state understood. As dendritic cells are the most important
and with therapy150–155. The vast majority of these studies cell type for initiating T cell responses in cancer, deci-
have focused on the tumour itself rather than assessing phering why dendritic cells are functionally impaired in
how the global immune macroenvironment is altered the periphery of individuals burdened with tumours is
compared with healthy individuals or how the periph- imperative. To date, studies have evaluated the capac-
eral immune landscape changes in response to ther- ity of tumour-bearing mice to mount type 1 immune
apy. A complete understanding of cancer and the host responses. As the immune system can be engaged in
immune responses across diverse tumour types, patient substantially different ways based on the context of the
populations and therapies requires detailed under- challenge, an intriguing next step is to drive function-
standing not only of the TME but also of the macro ally distinct immune responses in tumour-burdened
environmental alterations in immune organization. mice such as parasite and allergen challenges. Diverse
Unbiased single-cell technologies measuring the trans functional challenges will elucidate emergent func-
criptome, epigenome and proteome as well as multiple tional immunological alterations that are undetectable
modalities simultaneously will play an integral role in by simply examining cell population frequencies and
the construction of comprehensive organism-s cale phenotypes. Importantly, diverse functional challenges
reference maps of the immune system in cancer and may also reveal aspects of immunity that are not per-
of the impact of various cancer therapies. Establishing turbed by the tumour burden and, thus, inform thera-
distinct types of peripheral immune organization in peutic interventions that utilize unperturbed aspects of
patients with cancer will aid personalized medicine the immune system. Another important strategy is to
efforts by informing the context in which therapeutic design studies to thoroughly evaluate human immune
interventions will be introduced. Such studies in the responses to new challenges in patients with cancer
context of therapy will also inform how the peripheral who are treatment naive. Detailed single-cell analysis
0123456789();:
Reviews
of vaccine responses in patients with cancer as well as a susceptible population for infection, these studies can
longitudinal monitoring of patients with high-risk muta- inform vaccination formulations and therapeutic inter-
tions or organoid systems could be utilized to define the ventions to protect this vulnerable patient population
functionality of the tumour-burdened immune state from secondary infections. Altogether, an improved
in humans. These types of studies are an important understanding of how the peripheral immune land-
step towards identifying weaknesses, or perhaps new scape is perturbed and contributes to tumour control
strengths, in immune function that can rationally inform will provide essential next steps for the field.
the design and implementation of new cancer immuno-
therapies. Additionally, as patients with cancer represent Published online 9 April 2021
1. Coussens, L. M. & Werb, Z. Inflammation and cancer. 20. Shojaei, F. et al. G-CSF-initiated myeloid cell dramatically affects the differentiation of multiple
Nature 420, 860–867 (2002). mobilization and angiogenesis mediate tumor hematopoietic lineages in vivo. Blood 92, 4150–4166
2. Haslam, A. & Prasad, V. Estimation of the percentage refractoriness to anti-VEGF therapy in mouse models. (1998).
of US patients with cancer who are eligible for and Proc. Natl Acad. Sci. USA 106, 6742–6747 (2009). 38. Gabrilovich, D. I. et al. Production of vascular
respond to checkpoint inhibitor immunotherapy drugs. 21. Morales, J. K., Kmieciak, M., Knutson, K. L., endothelial growth factor by human tumors inhibits
JAMA Netw. Open 2, e192535 (2019). Bear, H. D. & Manjili, M. H. GM-CSF is one of the the functional maturation of dendritic cells. Nat. Med.
3. Hegde, P. S. & Chen, D. S. Top 10 challenges in cancer main breast tumor-derived soluble factors involved 2, 1096–1103 (1996).
immunotherapy. Immunity 52, 17–35 (2020). in the differentiation of CD11b–Gr1– bone marrow 39. Lin, J. H. et al. Type 1 conventional dendritic cells
4. Schultze, J. L., Mass, E. & Schlitzer, A. Emerging progenitor cells into myeloid-derived suppressor cells. are systemically dysregulated early in pancreatic
principles in myelopoiesis at homeostasis and during Breast Cancer Res. Treat. 123, 39–49 (2010). carcinogenesis. J. Exp. Med. 217, e201906733
infection and inflammation. Immunity 50, 288–301 22. Bayne, L. J. et al. Tumor-derived granulocyte– (2020).
(2019). macrophage colony-stimulating factor regulates This paper demonstrates that peripheral dendritic
5. Masopust, D. & Schenkel, J. M. The integration of myeloid inflammation and T cell immunity in cells are decreased via tumour-secreted IL-6
T cell migration, differentiation and function. Nat. Rev. pancreatic cancer. Cancer Cell 21, 822–835 (2012). inducing apoptosis. This deficit then leads to an
Immunol. 13, 309–320 (2013). 23. Raccosta, L. et al. The oxysterol–CXCR2 axis plays impaired peripheral response to vaccination that
6. Hanahan, D. & Coussens, L. M. Accessories to the a key role in the recruitment of tumor-promoting can be rescued by combination agonist anti-CD40
crime: functions of cells recruited to the tumor neutrophils. J. Exp. Med. 210, 1711–1728 (2013). and FLT3L treatment.
microenvironment. Cancer Cell 21, 309–322 (2012). 24. Dominguez, C., McCampbell, K. K., David, J. M. 40. Tjomsland, V. et al. Semi mature blood dendritic cells
7. Gonzalez, H., Hagerling, C. & Werb, Z. Roles of the & Palena, C. Neutralization of IL-8 decreases tumor exist in patients with ductal pancreatic adenocarcinoma
immune system in cancer: from tumor initiation to PMN-MDSCs and reduces mesenchymalization of owing to inflammatory factors released from the tumor.
metastatic progression. Genes Dev. 32, 1267–1284 claudin-low triple-negative breast cancer. JCI Insight 2, PLoS ONE 5, e13441 (2010).
(2018). e94296 (2017). 41. Dudek, A. M., Martin, S., Garg, A. D. & Agostinis, P.
8. Gabrilovich, D. I., Ostrand-Rosenberg, S. & Bronte, V. 25. Al Sayed, M. F. et al. T-cell-secreted TNFα induces Immature, semi-mature, and fully mature dendritic
Coordinated regulation of myeloid cells by tumours. emergency myelopoiesis and myeloid-derived cells: toward a DC-cancer cells interface that augments
Nat. Rev. Immunol. 12, 253–268 (2012). suppressor cell differentiation in cancer. Cancer Res. anticancer immunity. Front. Immunol. 4, 438 (2013).
9. Jaillon, S. et al. Neutrophil diversity and plasticity in 79, 346–359 (2019). 42. Ray-Coquard, I. et al. Lymphopenia as a prognostic
tumour progression and therapy. Nat. Rev. Cancer 20, 26. Peinado, H. et al. Melanoma exosomes educate bone factor for overall survival in advanced carcinomas,
485–503 (2020). marrow progenitor cells toward a pro-metastatic sarcomas, and lymphomas. Cancer Res. 69,
10. Canè, S. et al. The endless saga of monocyte diversity. phenotype through MET. Nat. Med. 18, 883–891 5383–5391 (2009).
Front. Immunol. 10, 1786 (2019). (2012). 43. Manuel, M. et al. Lymphopenia combined with low
11. Almand, B. et al. Increased production of immature 27. Wellenstein, M. D. et al. Loss of p53 triggers WNT- TCR diversity (divpenia) predicts poor overall survival
myeloid cells in cancer patients: a mechanism of dependent systemic inflammation to drive breast in metastatic breast cancer patients. Oncoimmunology
immunosuppression in cancer. J. Immunol. 166, cancer metastasis. Nature 572, 538–542 (2019). 1, 432–440 (2012).
678–689 (2001). 28. Allen, B. M. et al. Systemic dysfunction and plasticity 44. Liu, Y. Y. et al. Characteristics and prognostic
12. Casbon, A. J. et al. Invasive breast cancer reprograms of the immune macroenvironment in cancer models. significance of profiling the peripheral blood T-cell
early myeloid differentiation in the bone marrow to Nat. Med. 26, 1125–1134 (2020). receptor repertoire in patients with advanced lung
generate immunosuppressive neutrophils. Proc. Natl This paper demonstrates that the systemic cancer. Int. J. Cancer 145, 1423–1431 (2019).
Acad. Sci. USA 112, E566–E575 (2015). immune landscape in many mouse tumour models 45. Cui, J. H. et al. TCR repertoire as a novel indicator
13. Wu, W. C. et al. Circulating hematopoietic stem and is substantially perturbed yet can be restored for immune monitoring and prognosis assessment
progenitor cells are myeloid-biased in cancer patients. by tumour resection or cytokine treatment. of patients with cervical cancer. Front. Immunol. 9,
Proc. Natl Acad. Sci. USA 111, 4221–4226 (2014). Furthermore, new immune responses in the 2729 (2018).
This paper demonstrates that haematopoietic stem periphery are significantly blunted but can be 46. Tumeh, P. C. et al. PD-1 blockade induces responses
and progenitor cell populations are mobilized into rescued by agonist anti-CD40 treatment or surgical by inhibiting adaptive immune resistance. Nature 515,
the blood and biased towards myeloid differentiation resection. 568–571 (2014).
in multiple diverse human tumour types. 29. Hildner, K. et al. Batf3 deficiency reveals a critical role 47. Britanova, O. V. et al. Age-related decrease in TCR
14. Kersten, K. et al. Mammary tumor-derived CCL2 for CD8α+ dendritic cells in cytotoxic T cell immunity. repertoire diversity measured with deep and normalized
enhances pro-metastatic systemic inflammation Science 322, 1097–1100 (2008). sequence profiling. J. Immunol. 192, 2689–2698
through upregulation of IL1β in tumor-associated 30. Binnewies, M. et al. Unleashing type-2 dendritic cells (2014).
macrophages. Oncoimmunology 6, e1334744 to drive protective antitumor CD4+ T cell immunity. 48. Baum, P. D. et al. Blood T-cell receptor diversity
(2017). Cell 177, 556–571.e16 (2019). decreases during the course of HIV infection, but the
15. Coffelt, S. B. et al. IL-17-producing γδ T cells and 31. Mastelic-Gavillet, B. et al. Quantitative and qualitative potential for a diverse repertoire persists. Blood 119,
neutrophils conspire to promote breast cancer impairments in dendritic cell subsets of patients with 3469–3477 (2012).
metastasis. Nature 522, 345–348 (2015). ovarian or prostate cancer. Eur. J. Cancer 135, 49. Verronèse, E. et al. Immune cell dysfunctions in breast
16. Giles, A. J. et al. Activation of hematopoietic stem/ 173–182 (2020). cancer patients detected through whole blood multi-
progenitor cells promotes immunosuppression within 32. Della Bella, S. et al. Altered maturation of peripheral parametric flow cytometry assay. Oncoimmunology 5,
the pre-metastatic niche. Cancer Res. 76, 1335–1347 blood dendritic cells in patients with breast cancer. e1100791 (2015).
(2016). Br. J. Cancer 89, 1463–1723 (2003). 50. Wang, L. et al. IL6 signaling in peripheral blood T cells
17. Sio, A. et al. Dysregulated hematopoiesis caused by 33. Tabarkiewicz, J., Rybojad, P., Jablonka, A. & Rolinski, J. predicts clinical outcome in breast cancer. Cancer Res.
mammary cancer is associated with epigenetic changes CD1c+ and CD303+ dendritic cells in peripheral blood, 77, 1119–1126 (2017).
and Hox gene expression in hematopoietic cells. lymph nodes and tumor tissue of patients with 51. Wolf, A. M. et al. Increase of regulatory T cells in the
Cancer Res. 73, 5892–5904 (2013). non-small cell lung cancer. Oncol. Rep. 19, 237–243 peripheral blood of cancer patients. Clin. Cancer Res.
18. Meyer, M. A. et al. Breast and pancreatic cancer (2008). 9, 606–612 (2003).
interrupt IRF8-dependent dendritic cell development 34. Van Cruijsen, H. et al. Sunitinib-induced myeloid lineage 52. Ahmadzadeh, M. et al. Tumor-infiltrating human CD4+
to overcome immune surveillance. Nat. Commun. 9, redistribution in renal cell cancer patients: CD1c+ regulatory T cells display a distinct TCR repertoire and
1250 (2018). dendritic cell frequency predicts progression-free exhibit tumor and neoantigen reactivity. Sci. Immunol.
This paper demonstrates that disruption of survival. Clin. Cancer Res. 14, 5884–5892 (2008). 4, eaao4310 (2019).
haematopoiesis in human and mouse breast 35. Almand, B. et al. Clinical significance of defective 53. Wang, L. et al. Connecting blood and intratumoral
and pancreatic cancer decreases bone marrow dendritic cell differentiation in cancer. Clin. Cancer Res. Treg cell activity in predicting future relapse in breast
production of dendritic cells in the periphery. 6, 1755–1766 (2000). cancer. Nat. Immunol. 20, 1220–1230 (2019).
This deficit then results in impaired response to a 36. Failli, A., Legitimo, A., Orsini, G., Romanini, A. 54. Yuen, G. J., Demissie, E. & Pillai, S. B lymphocytes and
secondary peripheral challenge of a Matrigel plug & Consolini, R. Numerical defect of circulating cancer: a love–hate relationship. Trends Cancer 2,
containing OVA. dendritic cell subsets and defective dendritic cell 747–757 (2016).
19. Templeton, A. J. et al. Prognostic role of neutrophil- generation from monocytes of patients with advanced 55. Wang, W. et al. CD19+CD24hiCD38hi Bregs involved in
to-lymphocyte ratio in solid tumors: a systematic melanoma. Cancer Lett. 337, 184–192 (2013). downregulate helper T cells and upregulate regulatory
review and meta-analysis. J. Natl Cancer Inst. 106, 37. Gabrilovich, D. et al. Vascular endothelial growth T cells in gastric cancer. Oncotarget 6, 33486–33499
dju124 (2014). factor inhibits the development of dendritic cells and (2015).
0123456789();:
Reviews
56. Murakami, Y. et al. Increased regulatory B cells are immunity leading to postoperative cancer recurrence. 103. Scott, A. C. et al. TOX is a critical regulator of tumour-
involved in immune evasion in patients with gastric PLoS ONE 11, e0155947 (2016). specific T cell differentiation. Nature 571, 270–274
cancer. Sci. Rep. 9, 13083 (2019). 82. Tai, L. H. et al. Preventing postoperative metastatic (2019).
57. Zhou, J. et al. Enhanced frequency and potential disease by inhibiting surgery-induced dysfunction in 104. Khan, O. et al. TOX transcriptionally and epigenetically
mechanism of B regulatory cells in patients with lung natural killer cells. Cancer Res. 73, 97–107 (2013). programs CD8+ T cell exhaustion. Nature 571,
cancer. J. Transl Med. 12, 304 (2014). 83. Angka, L. et al. Natural killer cell IFNγ secretion is 211–218 (2019).
58. Olkhanud, P. B. et al. Tumor-evoked regulatory B cells profoundly suppressed following colorectal cancer 105. Yu, Y.-R. et al. Disturbed mitochondrial dynamics in
promote breast cancer metastasis by converting surgery. Ann. Surg. Oncol. 25, 3747–3754 (2018). CD8+ TILs reinforce T cell exhaustion. Nat. Immunol.
resting CD4+ T cells to T-regulatory cells. Cancer Res. 84. Danna, E. A. et al. Surgical removal of primary tumor 21, 1540–1551 (2020).
71, 3505–3515 (2011). reverses tumor-induced immunosuppression despite 106. Fourcade, J. et al. Upregulation of Tim-3 and PD-1
59. Shimasaki, N., Jain, A. & Campana, D. NK cells for the presence of metastatic disease. Cancer Res. 64, expression is associated with tumor antigen-specific
cancer immunotherapy. Nat. Rev. Drug Discov. 19, 2205–2211 (2004). CD8+ T cell dysfunction in melanoma patients. J. Exp.
200–218 (2020). 85. Mathios, D. et al. Anti-PD-1 antitumor immunity Med. 207, 2175–2186 (2010).
60. Mamessier, E. et al. Human breast cancer cells is enhanced by local and abrogated by systemic 107. Sade-Feldman, M. et al. Defining T cell states
enhance self tolerance by promoting evasion from chemotherapy in GBM. Sci. Transl Med. 8, 370ra180 associated with response to checkpoint immunotherapy
NK cell antitumor immunity. J. Clin. Invest. 121, (2016). in melanoma. Cell 175, 998–1013.e20 (2018).
3609–3622 (2011). This paper shows that systemic chemotherapy 108. Yost, K. E. et al. Clonal replacement of tumor-specific
61. Ménard, C. et al. Natural killer cell IFN-γ levels predict that disrupts systemic immune cell populations T cells following PD-1 blockade. Nat. Med. 25,
long-term survival with imatinib mesylate therapy in abrogates response to checkpoint blockade in 1251–1259 (2019).
gastrointestinal stromal tumor-bearing patients. glioblastoma. Conversely, local chemotherapy that This paper demonstrates that checkpoint blockade
Cancer Res. 69, 3563–3569 (2009). maintains peripheral immune integrity synergizes immunotherapy drives the expansion of novel
62. Charrier, M. et al. Circulating innate immune markers with checkpoint blockade. T cell clones within the tumour rather than the
and outcomes in treatment-naïve advanced non-small 86. Broz, M. L. et al. Dissecting the tumor myeloid reinvigoration of existing T cell clonotypes.
cell lung cancer patients. Eur. J. Cancer 108, 88–96 compartment reveals rare activating antigen- 109. Wu, T. D. et al. Peripheral T cell expansion predicts
(2019). presenting cells critical for T cell immunity. Cancer Cell tumour infiltration and clinical response. Nature 579,
63. Semeraro, M. et al. Clinical impact of the NKp30/ 26, 638–652 (2014). 274–278 (2020).
B7-H6 axis in high-risk neuroblastoma patients. This paper shows that dendritic cells acquire tumour This paper shows that expanded T cell clonotypes
Sci. Transl Med. 7, 283ra55 (2015). antigen within the TME, traffic to the tumour dLN after immunotherapy can be found in peripheral
64. Platonova, S. et al. Profound coordinated alterations and then prime antigen-specific T cells. blood including new T cell clones that infiltrate the
of intratumoral NK cell phenotype and function in lung 87. Roberts, E. W. et al. Critical role for CD103+/CD141+ tumour after treatment.
carcinoma. Cancer Res. 71, 5412–5422 (2011). dendritic cells bearing CCR7 for tumor antigen 110. Valpione, S. et al. Immune awakening revealed by
65. Shaked, Y. The pro-tumorigenic host response to trafficking and priming of T cell immunity in melanoma. peripheral T cell dynamics after one cycle of
cancer therapies. Nat. Rev. Cancer 19, 667–685 Cancer Cell 30, 324–336 (2016). immunotherapy. Nat. Cancer 1, 210–221 (2020).
(2019). 88. Salmon, H. et al. Expansion and activation of CD103+ 111. Kvistborg, P. et al. Anti-CTLA-4 therapy broadens the
66. Gustafson, C. E. et al. Immune cell repertoires in dendritic cell progenitors at the tumor site enhances melanoma-reactive CD8+ T cell response. Sci. Transl
breast cancer patients after adjuvant chemotherapy. tumor responses to therapeutic PD-L1 and BRAF Med. 6, 254ra128 (2014).
5, e134569 (2020). inhibition. Immunity 44, 924–938 (2016). 112. Kamphorst, A. O. et al. Rescue of exhausted CD8
67. Talebian Yazdi, M. et al. Standard radiotherapy but 89. Spranger, S., Dai, D., Horton, B. & Gajewski, T. F. T cells by PD-1-targeted therapies is CD28-dependent.
not chemotherapy impairs systemic immunity in non- Tumor-residing Batf3 dendritic cells are required for Science 355, 1423–1427 (2017).
small cell lung cancer. Oncoimmunology 5, e1255393 effector T cell trafficking and adoptive T cell therapy. 113. Hui, E. et al. T cell costimulatory receptor CD28 is a
(2016). Cancer Cell 31, 711–723.e4 (2017). primary target for PD-1-mediated inhibition. Science
68. van Meir, H. et al. Impact of (chemo)radiotherapy 90. Ruhland, M. K. et al. Visualizing synaptic transfer of 355, 1428–1433 (2017).
on immune cell composition and function in cervical tumor antigens among dendritic cells. Cancer Cell 37, 114. Ferrara, R. et al. Circulating T-cell immunosenescence
cancer patients. Oncoimmunology 6, e1267095 786–799.e5 (2020). in patients with advanced non-small cell lung cancer
(2017). 91. Chen, D. S. & Mellman, I. Oncology meets immunology: treated with single-agent PD-1/PD-L1 inhibitors or
69. Wesolowski, R. et al. Circulating myeloid-derived the cancer–immunity cycle. Immunity 39, 1–10 (2013). platinum-based chemotherapy. Clin. Cancer Res. 27,
suppressor cells increase in patients undergoing 92. Spitzer, M. H. et al. Systemic immunity is required for 492–503 (2021).
neo-adjuvant chemotherapy for breast cancer. Cancer effective cancer immunotherapy. Cell 168, 487–502 115. Cader, F. Z. et al. A peripheral immune signature
Immunol. Immunother. 66, 1437–1447 (2017). (2017). of responsiveness to PD-1 blockade in patients
70. Larsson, A. M., Roxå, A., Leandersson, K. This paper demonstrates that trafficking of immune with classical Hodgkin lymphoma. Nat. Med. 26,
& Bergenfelz, C. Impact of systemic therapy on cells is required for effective immunotherapy. 1468–1479 (2020).
circulating leukocyte populations in patients with Furthermore, a peripheral signature of CD4+ T cells 116. Vonderheide, R. H. CD40 agonist antibodies in cancer
metastatic breast cancer. Sci. Rep. 9, 13451 (2019). expressing low levels of PD1 and CD127 marked immunotherapy. Annu. Rev. Med. 71, 47–58 (2020).
71. Valdés-Ferrada, J. et al. Peripheral blood classical patients who responded to immunotherapy. 117. Li, D.-K. & Wang, W. Characteristics and clinical trial
monocytes and plasma interleukin 10 are associated 93. Fransen, M. F. & Van Hall, T. Tumor-draining lymph results of agonistic anti-CD40 antibodies in the
to neoadjuvant chemotherapy response in breast nodes are pivotal in PD-1/PD-L1 checkpoint therapy. treatment of malignancies (Review). Oncol. Lett. 20,
cancer patients. Front. Immunol. 11, 1413 (2020). JCI Insight 3, e124507 (2018). 1–1 (2020).
72. Axelrod, M. L. et al. Changes in peripheral and local 94. Lau, J. et al. Tumour and host cell PD-L1 is required to 118. Morrison, A. H., Diamond, M. S., Hay, C. A.,
tumor immunity after neoadjuvant chemotherapy mediate suppression of anti-tumour immunity in mice. Byrne, K. T. & Vonderheide, R. H. Sufficiency of CD40
reshape clinical outcomes in patients with breast Nat. Commun. 8, 14572 (2017). activation and immune checkpoint blockade for T cell
cancer. Clin. Cancer Res. 26, 5668–5681 (2020). 95. Lin, H. et al. Host expression of PD-L1 determines priming and tumor immunity. Proc. Natl Acad. Sci.
73. Yu, W. D., Sun, G., Li, J., Xu, J. & Wang, X. efficacy of PD-L1 pathway blockade-mediated tumor USA 117, 8022–8031 (2020).
Mechanisms and therapeutic potentials of cancer regression. J. Clin. Invest. 128, 805–815 (2018). 119. O’Hara, M. H. et al. CD40 agonistic monoclonal
immunotherapy in combination with radiotherapy 96. Strauss, L. et al. Targeted deletion of PD-1 in myeloid antibody APX005M (sotigalimab) and chemotherapy,
and/or chemotherapy. Cancer Lett. 452, 66–70 (2019). cells induces antitumor immunity. Sci. Immunol. 5, with or without nivolumab, for the treatment of
74. Luo, Q., Zhang, L., Luo, C. & Jiang, M. Emerging eaay1863 (2020). metastatic pancreatic adenocarcinoma: an open-label,
strategies in cancer therapy combining chemotherapy 97. Oh, S. A. et al. PD-L1 expression by dendritic cells is a multicentre, phase 1b study. Lancet Oncol. 22,
with immunotherapy. Cancer Lett. 454, 191–203 key regulator of T-cell immunity in cancer. Nat. Cancer 118–131 (2021).
(2019). 1, 681–691 (2020). 120. Osborne, L. C. et al. Virus–helminth coinfection
75. Bailly, C., Thuru, X. & Quesnel, B. Combined cytotoxic This paper shows that deletion of PDL1 specifically reveals a microbiota-independent mechanism of
chemotherapy and immunotherapy of cancer: modern in dendritic cells, but not macrophages, greatly immunomodulation. Science 345, 578–582 (2014).
times. NAR Cancer 2, zcaa002 (2020). restricts tumour growth and amplifies antitumour 121. Barnstorf, I. et al. Chronic virus infection compromises
76. Tohme, S., Simmons, R. L. & Tsung, A. Surgery for CD8+ T cell responses. memory bystander T cell function in an IL-6/STAT1-
cancer: a trigger for metastases. Cancer Res. 77, 98. Dammeijer, F. et al. The PD-1/PD-L1-checkpoint dependent manner. J. Exp. Med. 216, 571–586
1548–1552 (2017). restrains T cell immunity in tumor-draining lymph (2019).
77. Tohme, S. et al. Neutrophil extracellular traps promote nodes. Cancer Cell 38, 1–16 (2020). 122. Snell, L. M. et al. CD8+ T cell priming in established
the development and progression of liver metastases 99. Peng, Q. et al. PD-L1 on dendritic cells attenuates chronic viral infection preferentially directs
after surgical stress. Cancer Res. 76, 1367–1380 T cell activation and regulates response to immune differentiation of memory-like cells for sustained
(2016). checkpoint blockade. Nat. Commun. 11, 4835 (2020). immunity. Immunity 49, 678–694.e5 (2018).
78. Krall, J. A. et al. The systemic response to surgery 100. Mayoux, M. et al. Dendritic cells dictate responses to 123. Pollyea, D. A., Brown, J. M. Y. & Horning, S. J. Utility
triggers the outgrowth of distant immune-controlled PD-L1 blockade cancer immunotherapy. Sci. Transl of influenza vaccination for oncology patients. J. Clin.
tumors in mouse models of dormancy. Sci. Transl Med. Med. 12, eaav7431 (2020). Oncol. 28, 2481–2490 (2010).
10, eaan3464 (2018). 101. Chamoto, K. et al. Mitochondrial activation chemicals 124. Liang, W. et al. Cancer patients in SARS-CoV-2
79. Bosiljcic, M. et al. Targeting myeloid-derived suppressor synergize with surface receptor PD-1 blockade for infection: a nationwide analysis in China. Lancet Oncol.
cells in combination with primary mammary tumor T cell-dependent antitumor activity. Proc. Natl Acad. 21, 335–337 (2020).
resection reduces metastatic growth in the lungs. Sci. USA 114, E761–E770 (2017). 125. Kuderer, N. M. et al. Clinical impact of COVID-19 on
Breast Cancer Res. 21, 103 (2019). 102. Philip, M. et al. Chromatin states define tumour- patients with cancer (CCC19): a cohort study. Lancet
80. Kallis, M. P. et al. Pharmacological prevention of specific T cell dysfunction and reprogramming. Nature 395, 1907–1918 (2020).
surgery-accelerated metastasis in an animal model 545, 452–456 (2017). 126. Mittal, R. et al. Phenotypic T cell exhaustion in a
of osteosarcoma. J. Transl. Med. 18, 183 (2020). This paper shows that intratumoural T cells acquire murine model of bacterial infection in the setting
81. Ananth, A. A. et al. Surgical stress abrogates a terminal dysfunctional state that can no longer of pre-existing malignancy. PLoS ONE 9, e93523
pre-existing protective T cell mediated anti-tumor participate in tumour clearance. (2014).
0123456789();:
Reviews
127. Russ, A. J. et al. Melanoma-induced suppression of checkpoint blockade in patients with metastatic immunosuppressive CD14+HLA–DR–/low cells in
tumor antigen-specific T cell expansion is comparable melanoma. Nat. Med. 26, 193–199 (2020). melanoma patients are Stat3hi and overexpress
to suppression of global T cell expansion. Cell. Immunol. This paper finds that patients with metastatic CD80, CD83, and DC-sign. Cancer Res. 70,
271, 104–109 (2011). melanoma responding to checkpoint blockade show 4335–4345 (2010).
128. Adashek, J., Goloubev, A., Kato, S. & Kurzrock, R. clonal expansion of CD8+ T cells with a cytotoxic 168. Santegoets, S. J. et al. Myeloid derived suppressor
Immunotherapy trials lack a biomarker for inclusion: phenotype in peripheral blood. and dendritic cell subsets are related to clinical
implications for drug development. J. Immunother. 148. Koh, J. et al. Regulatory (FoxP3+) T cells and TGF-β outcome in prostate cancer patients treated with
Cancer 8, A192 (2020). predict the response to anti-PD-1 immunotherapy in prostate GVAX and ipilimumab. J. Immunother. Cancer
129. Hardy-Werbin, M. et al. Serum cytokine levels as patients with non-small cell lung cancer. Sci. Rep. 10, 2, 31 (2014).
predictive biomarkers of benefit from ipilimumab in 18994 (2020). 169. Meloni, F. et al. Foxp3 expressing CD4+CD25+ and
small cell lung cancer. Oncoimmunology 8, 1593810 149. Kim, H. R. et al. The ratio of peripheral regulatory CD8+CD28– T regulatory cells in the peripheral blood
(2019). T cells to Lox-1 1 polymorphonuclear myeloid-derived of patients with lung cancer and pleural mesothelioma.
130. Boutsikou, E. et al. Tumour necrosis factor, suppressor cells predicts the early response to anti- Hum. Immunol. 67, 1–12 (2006).
interferon-γ and interleukins as predictive markers PD-1 therapy in patients with non-small cell lung 170. Miller, A. M. et al. CD4+CD25high T cells are enriched in
of antiprogrammed cell-death protein-1 treatment cancer. Am. J. Respir Crit. Care Med. 199, 243–246 the tumor and peripheral blood of prostate cancer
in advanced non-small cell lung cancer: a pragmatic (2019). patients. J. Immunol. 177, 7398–7405 (2006).
approach in clinical practice. Ther. Adv. Med. Oncol. 150. Tirosh, I. et al. Dissecting the multicellular ecosystem 171. Maruyama, T. et al. Distribution of TH17 cells and
10, 1758835918768238 (2018). of metastatic melanoma by single-cell RNA-seq. FoxP3+ regulatory T cells in tumor-infiltrating
131. Sanmamed, M. F. et al. Changes in serum Science 352, 189–196 (2016). lymphocytes, tumor-draining lymph nodes and
interleukin-8 (IL-8) levels reflect and predict response 151. Gubin, M. M. et al. High-dimensional analysis peripheral blood lymphocytes in patients with gastric
to anti-PD-1 treatment in melanoma and non-small- delineates myeloid and lymphoid compartment cancer. Cancer Sci. 101, 1947–1954 (2010).
cell lung cancer patients. Ann. Oncol. 28, 1988–1995 remodeling during successful immune-checkpoint 172. Ichihara, F. et al. Increased populations of regulatory
(2017). cancer therapy. Cell 175, 1014–1030.e19 (2018). T cells in peripheral blood and tumor-infiltrating
132. Weide, B. et al. Baseline biomarkers for outcome 152. Azizi, E. et al. Single-cell map of diverse immune lymphocytes in patients with gastric and esophageal
of melanoma patients treated with pembrolizumab. phenotypes in the breast tumor microenvironment. cancers. Clin. Cancer Res. 9, 4404–4408 (2003).
Clin. Cancer Res. 22, 5487–54967 (2016). Cell 174, 1293–1308 (2018). 173. Ormandy, L. et al. Increased populations of
133. Mezquita, L. et al. Association of the lung immune 153. Zilionis, R. et al. Single-cell transcriptomics of human regulatory T cells in peripheral blood of patients
prognostic index with immune checkpoint inhibitor and mouse lung cancers reveals conserved myeloid with hepatocellular carcinoma. Cancer Res. 65,
outcomes in patients with advanced non-small cell populations across individuals and species. Immunity 2457–2464 (2005).
lung cancer. JAMA Oncol. 4, 351–357 (2018). 50, 1317–1334.e10 (2019). 174. Liyanage, U. K. et al. Prevalence of regulatory
134. Hogan, S. A. et al. Peripheral blood TCR repertoire 154. Wagner, J. et al. A single-cell atlas of the tumor and T cells is increased in peripheral blood and tumor
profiling may facilitate patient stratification for immune ecosystem of human breast cancer. Cell 177, microenvironment of patients with pancreas or breast
immunotherapy against melanoma. Cancer Immunol. 1–16 (2019). adenocarcinoma. J. Immunol. 169, 2756–2761
Res. 7, 77–85 (2019). 155. Steele, N. G. et al. Multimodal mapping of the tumor (2002).
135. Howard, R., Kanetsky, P. A. & Egan, K. M. Exploring and peripheral blood immune landscape in human 175. Schluter, J. et al. The gut microbiota is associated
the prognostic value of the neutrophil-to-lymphocyte pancreatic cancer. Nat. Cancer 1, 1097–1112 (2020). with immune cell dynamics in humans. Nature 588,
ratio in cancer. Sci. Rep. 9, 1–10 (2019). 156. Nabet, B. Y. et al. Noninvasive early identification of 303–307 (2020).
136. Capone, M. et al. Baseline neutrophil-to-lymphocyte therapeutic benefit from immune checkpoint 176. Routy, B. et al. Gut microbiome influences efficacy of
ratio (NLR) and derived NLR could predict overall inhibition. Cell 183, 363–376.e13 (2020). PD-1-based immunotherapy against epithelial tumors.
survival in patients with advanced melanoma 157. O’Hara, M. H. et al. A phase Ib study of CD40 Science 359, 91–97 (2018).
treated with nivolumab. J. Immunother. Cancer 6, 74 agonistic monoclonal antibody APX005M together 177. Gopalakrishnan, V. et al. Gut microbiome modulates
(2018). with gemcitabine (Gem) and nab-paclitaxel (NP) with response to anti-PD-1 immunotherapy in melanoma
137. Lalani, A.-K. A. et al. Change in neutrophil- or without nivolumab (Nivo) in untreated metastatic patients. Science 359, 97–103 (2018).
to-lymphocyte ratio (NLR) in response to immune ductal pancreatic adenocarcinoma (PDAC) patients. 178. Matson, V. et al. The commensal microbiome is
checkpoint blockade for metastatic renal cell Cancer Res. 79, CT004 (2019). associated with anti-PD-1 efficacy in metastatic
carcinoma. J. Immunother. Cancer 6, 7 (2018). 158. Kyi, C. et al. Therapeutic immune modulation against melanoma patients. Science 359, 104–108 (2018).
138. Ferrucci, P. F. et al. Baseline neutrophils and derived solid cancers with intratumoral poly-ICLC: a pilot trial. 179. Mager, L. F. et al. Microbiome-derived inosine
neutrophilto-lymphocyte ratio: prognostic relevance Clin. Cancer Res. 24, 4937–4948 (2018). modulates response to checkpoint inhibitor
in metastatic melanoma patients receiving ipilimumab. 159. Bhardwaj, N. et al. Flt3 ligand augments immune immunotherapy. Science 369, 1481–1489 (2020).
Ann. Oncol. 27, 732–738 (2016). responses to anti-DEC-205-NY-ESO-1 vaccine through 180. Tanoue, T. et al. A defined commensal consortium
139. Kitano, S. et al. Computational algorithm-driven expansion of dendritic cell subsets. Nat. Cancer 1, elicits CD8 T cells and anti-cancer immunity. Nature
evaluation of monocytic myeloid-derived suppressor 1204–1217 (2020). 565, 600–605 (2019).
cell frequency for prediction of clinical outcomes. 160. Brandau, S. et al. Myeloid-derived suppressor cells 181. Fessler, J., Matson, V. & Gajewski, T. F. Exploring
Cancer Immunol. Res. 2, 812–821 (2014). in the peripheral blood of cancer patients contain the emerging role of the microbiome in cancer
140. Martens, A. et al. Baseline peripheral blood a subset of immature neutrophils with impaired immunotherapy. J. Immunother. Cancer 7, 108 (2019).
biomarkers associated with clinical outcome migratory properties. J. Leukoc. Biol. 89, 311–317
of advanced melanoma patients treated with (2011). Acknowledgements
ipilimumab. Clin. Cancer Res. 22, 2908–2918 161. Raychaudhuri, B. et al. Myeloid-derived suppressor The authors acknowledge funding support from National
(2016). cell accumulation and function in patients with newly Institutes of Health (NIH) grant DP5 OD023056, the Parker
141. Gebhardt, C. et al. Myeloid cells and related chronic diagnosed glioblastoma. Neuro. Oncol. 13, 591–599 Institute for Cancer Immunotherapy and the Chan Zuckerberg
inflammatory factors as novel predictive markers in (2011). Biohub to M.H.S.
melanoma treatment with ipilimumab. Clin. Cancer 162. Zhu, Y. P. et al. Identification of an early unipotent
Res. 21, 5453–5459 (2015). neutrophil progenitor with pro-tumoral activity in Author contributions
142. Meyer, C. et al. Frequencies of circulating MDSC mouse and human bone marrow. Cell Rep. 24, The authors contributed equally to all aspects of the article.
correlate with clinical outcome of melanoma patients 2329–2341.e8 (2018).
treated with ipilimumab. Cancer Immunol. Immunother. 163. Sagiv, J. Y. et al. Phenotypic diversity and plasticity Competing interests
63, 247–257 (2014). in circulating neutrophil subpopulations in cancer. M.H.S. is a co-founder and board member of Teiko Bio,
143. Retseck, J. et al. Long term impact of CTLA4 blockade Cell Rep. 10, 562–573 (2015). receives research funding from Roche/Genentech, Bristol-
immunotherapy on regulatory and effector immune 164. Chittezhath, M. et al. Molecular profiling reveals a Myers Squibb and Valitor, and has been a paid consultant for
responses in patients with melanoma. J. Transl Med. tumor-promoting phenotype of monocytes and Earli, Five Prime Therapeutics, Ono Pharmaceutical and
16, 184 (2018). macrophages in human cancer progression. Immunity January Inc. B.M.A. is currently an employee of Teiko Bio.
144. Mueller, S. et al. Mass cytometry detects 41, 815–294 (2014). K.J.H.-G. declares no competing interests.
H3.3K27M-specific vaccine responses in diffuse 165. Filipazzi, P. et al. Identification of a new subset
midline glioma. J. Clin. Invest. 130, 6325–6337 of myeloid suppressor cells in peripheral blood of Peer review information
(2020). melanoma patients with modulation by a granulocyte– Nature Reviews Cancer thanks N. Chaput, T. Merghoub and
145. Kotsakis, A. et al. Prognostic value of circulating macrophage colony-stimulation factor-based antitumor the other, anonymous, reviewer(s) for their contribution to the
regulatory T cell subsets in untreated non-small cell vaccine. J. Clin. Oncol. 25, 2546–2553 (2007). peer review of this work.
lung cancer patients. Sci. Rep. 6, 39247 (2016). 166. Hoechst, B. et al. A new population of myeloid-derived
146. Huang, A. C. et al. T-cell invigoration to tumour burden suppressor cells in hepatocellular carcinoma patients Publisher’s note
ratio associated with anti-PD-1 response. Nature 545, induces CD4+CD25+Foxp3+ T cells. Gastroenterology Springer Nature remains neutral with regard to jurisdictional
60–65 (2017). 135, 234–243 (2008). claims in published maps and institutional affiliations.
147. Fairfax, B. P. et al. Peripheral CD8+ T cell characteristics 167. Poschke, I., Mougiakakos, D., Hansson, J.,
associated with durable responses to immune Masucci, G. V. & Kiessling, R. Immature © Springer Nature Limited 2021
0123456789();:

Reviews: Systemic Immunity in Cancer

Uploaded by

Copyright:

Available Formats

Reviews: Systemic Immunity in Cancer

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Reviews: Systemic Immunity in Cancer

Uploaded by

Copyright:

Available Formats

REVIEWS

Systemic immunity in cancer

Peripheral immune cells

nATure RevIewS | CanceR volume 21 | June 2021 | 345

346 | June 2021 | volume 21 www.nature.com/nrc

b ↑ Frequencies ↓ Frequencies Altered function

Blood PMA and

c Spleen d Tumour dLN

nATure RevIewS | CanceR volume 21 | June 2021 | 347

348 | June 2021 | volume 21 www.nature.com/nrc

Table 1 | Peripheral immune perturbations in cancer

cells oesophageal cancer, lung cancer, ovarian cancer,

progenitor cells oesophageal cancer, lung cancer, ovarian cancer,

monocyte oesophageal cancer, lung cancer, ovarian cancer,

precursors Mouse: breast cancer, pancreatic cancer

neutrophils/ cancerM/NM, bladder cancerM/NM, head and neck 28,160–163

PMN-​MDSCs cancerM/NM, glioblastoma, melanoma

monocytes/M-​MDSCs cancer, melanomaM/NM, hepatocellular carcinoma 164–167

Mouse: breast cancer , melanoma , pancreatic

cancerM, colon cancerNM, glioblastoma

head and neck squamous cell carcinoma, melanomaM/NM, 32–34,36,168

lung cancerNM, renal cancer

cancerM/NM, colorectal cancer, oesophageal cancer, 169–174

hepatocellular carcinoma, pancreatic cancer, breast

cancer, breast cancer

Mouse: breast cancer M

T cells production after PMA and

pSTAT3 signalling after

receptors, increased gastrointestinal cancerM/NM, neuroblastomaM

nATure RevIewS | CanceR volume 21 | June 2021 | 349

350 | June 2021 | volume 21 www.nature.com/nrc

nATure RevIewS | CanceR volume 21 | June 2021 | 351

a Baseline treatment-naive tumour dLN

PD1 CD86 CD40

New active Dysfunctional

352 | June 2021 | volume 21 www.nature.com/nrc

nATure RevIewS | CanceR volume 21 | June 2021 | 353

a Healthy control Tumour-bearing mouse T cell response to L. monocytogenes in Pan02 pancre-

of OVA-​specific CD8+ T cells infiltrating the plug when

CD80 Systemic immune biomarkers for cancer

Despite significant interest in the development of predic-

help guide patient treatment decisions.

Circulating protein biomarkers. Quantification of cir-

354 | June 2021 | volume 21 www.nature.com/nrc

granzyme B+ at baseline (anti-​PD1) lymphoma defined as CD45RA−FOXP3hi were predictive of relapse

treatment (anti-​PD1) of mature NK cells and a subset of cells defined by

nATure RevIewS | CanceR volume 21 | June 2021 | 355

356 | June 2021 | volume 21 www.nature.com/nrc

nATure RevIewS | CanceR volume 21 | June 2021 | 357

358 | June 2021 | volume 21 www.nature.com/nrc

nATure RevIewS | CanceR volume 21 | June 2021 | 359

You might also like

PMN-MDSCs cancerM/NM, glioblastoma, melanoma

monocytes/M-MDSCs cancer, melanomaM/NM, hepatocellular carcinoma 164–167

of OVA-specific CD8+ T cells infiltrating the plug when

granzyme B+ at baseline (anti-PD1) lymphoma defined as CD45RA−FOXP3hi were predictive of relapse

treatment (anti-PD1) of mature NK cells and a subset of cells defined by