Seminars in Cancer Biology 42 (2017) 13–19

Contents lists available at ScienceDirect

Seminars in Cancer Biology

journal homepage: www.elsevier.com/locate/semcancer

Review

Roles of tumor heterogeneity in the development of drug resistance:

A call for precision therapy
Duojiao Wu ∗ , Diane C. Wang, Yunfeng Cheng, Mengjia Qian, Miaomiao Zhang, Qi Shen,
Xiangdong Wang
Zhongshan Hospital Institute of Clinical Science, Fudan University, Shanghai Institute of Clinical Bioinformatics, Shanghai, China

a r t i c l e i n f o a b s t r a c t

Article history: The drug resistance limits the optimal efficacy of drugs during target therapies for lung cancer and
Received 8 June 2016 requires the development of precision medicine to identify and develop new highly selective drugs
Accepted 8 November 2016 and more precise tailoring of medicine to the target population. Lung cancer heterogeneity as a poten-
Available online 10 November 2016
tial cause of drug resistance to targeted therapy may foster tumor evolution and adaptation and fade
personalized-medicine strategies. The present review elucidates the influence of tumor heterogeneity
Keywords:
on drug efficacy and resistance, and discusses potential strategies to combat heterogeneity for cancer
Tumor heterogeneity
treatment. There is an urgent need to discover and develop disease- and biology-specific biomarkers
Precision medicine
Genetic variations
for monitoring the existence and occurrence of lung cancer heterogeneity, testing targeted drugs in
Cancer clinical trials, and implementing precision medicine for patients. Better understanding of lung cancer
Drug resistance heterogeneity will strengthen therapeutic strategies and apply precision medicine to cure the disease.
Biomarker © 2016 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2. Disease heterogeneity and drug efficacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3. Micro-environment heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4. Metastatic heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5. Functional heterogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
6. Clinical potentials of precision medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
7. Functional and precision biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
8. Opportunities and challenges of precision therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Conflict of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

and variants, or protein over-expression [4]. Precision medicine

1. Introduction requires an understanding of cancer genes, mutational processes,
or heterogeneity between cancer cells during tumor evolution.
Precision medicine is proposed as a new strategy to identify and The tumor heterogeneity in diverse cancer types was evidenced
develop new highly selective drugs against specific targets for the by a meta-analysis of 2957 whole exomes and 126 whole genomes
disease and more precise tailoring of medicines to the target popu- [5]. Inter-tumor heterogeneity with limited somatic alterations was
lations [1–3]. Precision medicine can be an important approach to noticed in histopathologic subtype tumors, while intra-tumor het-
create more novel and safer therapeutics for patients with gene erogeneity within individual tumor biopsies of the same tumor and
fusions and mutation, methylation and acetylation, aberrations temporally evolves during the disease course [6–8].
Patients with advanced solid tumors still have poor clinical out-
come, due to the resistance to chemotherapeutics and targeted
∗ Corresponding author. therapies [9].The tumor heterogeneity may exist between individ-
E-mail address: wuduojiao@126.com (D. Wu). uals, organs/tissues, locations, or cells [10,11] and contribute to

http://dx.doi.org/10.1016/j.semcancer.2016.11.006
1044-579X/© 2016 Elsevier Ltd. All rights reserved.
14 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19

low efficacy or failure of therapies through the development of
drug resistance [12,13]. Acquired resistance could develop after
an initial phase and be assessed by clinical phenotypes [14,15].
Resistance mechanisms were involved in secondary pathway
mutations or bypass mechanisms within the tumor cells, such as
EGFR(T790M)44 mutations or MET receptor amplification [16,17].
Understanding of lung cancer heterogeneity can help the develop-
ment of predictive or prognostic biomarker strategies [18–21], to
monitor the influence of tumor subclones in therapeutic efficacies
and outcomes.
The present review briefly define various heterogeneities and
potential associations with drug efficacy and resistance. We then
overview new strategies to increase drug efficacy and minimal-
ize the drug resistance and toxicity. We emphasize the importance
to develop functional and precision biomarkers to monitor drug
efficacy and resistance, and define opportunities and challenges of
precision medicine for clinical practice.
2. Disease heterogeneity and drug efficacy
Tumor heterogeneity exists and develops between patients with
or without genetic factors. A sequencing analysis of 3281 samples
from 12 cancer types demonstrated that acute myeloid leukaemia
was associated with the lowest mutational burden, while lung
squamous cell carcinoma with the highest mutational burden
[22,23]. Some somatic mutations e.g. methylmalonyl-CoA mutase
−driver genes, caused interpatient tumor heterogeneity [24,25].
The high mutational burden of lung cancer samples was explained
by the mutagenic effect of carcinogen exposures to e.g. cigarette
smoke, and melanoma to ultraviolet radiation [26].
Lung cancer is a molecularly heterogeneous disease. The muta-
tional landscape of lung adenocarcinoma is substantially different
from that of squamous cell carcinoma or small cell lung cancer
(SCLC), as shown in Table 1. About 10%-40% mutations in the
tyrosine kinase domain of the epidermal growth factor recep-
tor (EGFR) were noticed in lung adenocarcinomas [22,27], while
rarely in squamous cell carcinoma and SCLC [22]. A number of
TP53, KRAS, STK11, EGFR, and NF1 were mutated in adenocarci-
nomas, different from those classified by smoke status and gender.
Mutations inTP53, KRAS, LKB1, NF1, and RBM10 were enriched in
transversion-high tumors, while EGFR, RB1, and PIK3CA, and in-
frame insertions in the receptor tyrosine kinases EGFR and ERBB2
in transversion-low tumors. Mutations in EGFR are associated with
women, while RBM10 with men. Host genetic or non-genetic fac-
tors determine drug half-life, vascular permeability to drugs or
acquired resistance to targeted drugs [28]. Cancers with low genetic
mutational burden are associated with long lasting response to
targeted therapy, whereas gene-unstable cancers with shorter
duration of responses. The smoking-associated cancers with high
levels of carcinogen-associated genetic mutations may be less to
derive substantial long-term benefit from targeted therapy.
3. Micro-environment heterogeneity
The environment and/or microenvironment where the human
and tumor cells live was recently identified as a factor to influence
drug resistance and emphasized as the importance of the tumor
cell extrinsic compartments [29,30]. The origin and influence of
the micro-environment heterogeneity were involved in the recruit-
ment of fibroblasts, migration of immune cells, matrix remodeling,
and development of vascular networks [31–34]. Phenotypic and
functional heterogeneity of cancer-associated fibroblasts could
promote tumorigenesis, extracellular matrix production, and
cytokine secretion, including stromal cell-derived factor 1,vascu-
lar endothelial growth factor, platelet-derived growth factor, and
hepatocyte growth factor (HGF) [35]. Cancer-associated stromal
secretion of HGF could activate the HGF receptor MET, mitogen-
activated protein kinase (MAPK), and phosphatidylinositol-3-OH
kinase (PI3K/AKT) signaling pathways, to have the immediate
resistance to RAF inhibition [31]. The activation of MAPK and
PI3 K/AKT-mTOR pathways regulates the formation of the eIF4F
eukaryotic translation complex in BRAF(V600)-mutant tumors.
The persistent formation of the eIF4F complex, binding to the
7-methylguanylate cap (m(7)G) at the 5’ end of messenger RNA
and modulating the translation of specific mRNAs, was associated
with resistance to anti-BRAF, anti-MEK, and anti-BRAF plus anti-
MEK drug combinations in BRAF(V600)-mutant melanoma, colon
and thyroid cancer cells [33]. Microvessel density was reported
to be a significant prognostic factor for poor outcome in NSCLC
[36]. The expression of vascular endothelial growth factor-A was
up-regulated with a worse prognosis in lung and renal cancers
[37]. Tumor infiltrating lymphocytes may recognize neo-antigens
presented on the surface of tumor cells as non-self, promoting
enhanced T cell activation and immune cell tumor infiltration [38].
T cell activation was involved with stimulatory and inhibitory
checkpoint signals to precisely tune responses to prevent excessive
damage and autoimmunity. The usurping cytotoxic T cells can be
activated in tumors through continuous engagement of inhibitory

Table 1
Heterogeneity of lung cancer.

Features Adenocarcinoma Squamous cell carcinoma Small-cell lung cancer

Age-adjusted incidence (Incidence per 100,000 per year) 22.1 14.4 9.8
Napsin-A and TTF-1 immunostaining Napsin-A (+) Napsin-A (−) Napsin-A (−)
TTF-1 (+) TTF-1 (−) TTF-1 (+)
Genomic lateration Mutations TP53, KRAS, EGFR, NF1, BRAF, TP53, CDKN2A, PIK3CA, TP53, RB1, EP300, CREBBP,
MET, RIT NFE2L2, KEAP1, PTEN, SLIT2
CUL3, PTEN, NF1, NOTCH1,2,
and 3, DDR2,
EGFR
Fusions ALK, ROS1, FGFRs
RET,NTRK1,RASSF1A,FZR2
Somatic copy number alteration Gains: NKX2-1, TERT, EGFR, Gains: Chr 3q 26 (SOX2, Gains: MYC, MYCN, MYCL1,
MET, KRAS,ERBB2, MDM2 PIK3CA, TP63 etc) SOX2, FGFR1, KIT
Losses: LRP1B, PTPRD, and Losses: CDKN2A, PTEN Losses: Chr 3p (FHIT, FUS1,
CDKN2A RASSF1A)
Pathway alterations RTK/RAS/RAF Squamous diff erentiation Hedgehog, DNA repair,axonal
mTOR Oxidative stress response guidance and neuroendocrine
JAK-STAT PIK3CA diff erentiation, Cell cycle
DNA repair DNA repair regulation, epigenetic
Cell cycle regulation Cell cycle regulation deregulation,
Epigenetic deregulation Epigenetic deregulation
 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19 15

Fig. 1. Interplay of tumor and microenvironment and the therapies of targetting the tumor microenvironment.
(A) Origins and influence of tumor local heterogeneity. Tumor formation involves the co-evolution of neoplastic cells together with extracellular matrix and vascular
endothelial, stromal and immune cells. Several chemokines secreted by cancer-associated fibroblasts(CAFs) modulate the stromal landscape: CXC-chemokine ligand 12
(CCL12) reples T cells, CXCL13 recruits B cells, and CC-chemokine ligand 2 (CCL2), CCL3, CCL4 and CCL5 recruit myeloid cells. Tumor-associated blood endothelial cells(BECs)
upregulated mucosal vascular addressin cell adhension molecule 1(MADCAM1) and CD166 expression, which bind to ␣4␤7 integrin and CD6, respectively, on regulatory T
(Treg) cells and promote their preferential extravasation through the activation of ␣4␤7 integrin. (B) The FDA-approved agents that target the tumor vascular network and
tumor infiltrating immune cells have been listed in the table of Fig. 1B. CSF1R, colony stimulating factor 1 receptor; CTLA-4, cytotoxic T-lymphocyte-associated antigen 4;
DC, dendritic cell; EGFR, epidermal growth factor receptor; FDA, Food and Drug Administration; FLT3, Fms-like tyrosine kinase 3; IFN, interferon; IL-2, interleukin 2; PAP,
prostatic acid phosphatase; PD-1, programmed death-1; PDGFR, Platelet-derived growth factor receptor; PDL1, programmed death ligand 1; PLGF, placental growth factor;
VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor.

receptors, e.g. cytotoxic T-lymphocyte-associated antigen-4 and
programmed death 1 by upregulation of their ligands [39].
The complexity of the tumor results from continuous crosstalk
between tumor cells and the environment, as explained in Fig. 1A,
in addition to the complication of temporal heterogeneity on top of
spatial heterogeneity. The Food and Drug Administration-approved
therapies targeting multiple compartments of micro-environment
has been listed in Fig. 1B. The question is whether durable responses
result from multi-pronged approaches if tumor cells and their cell-
extrinsic support can be targeted. It is still a challenge to define
the relationship between tumor heterogeneity and therapeutic
response.
4. Metastatic heterogeneity
The heterogeneity exists between the origin and metastatic
locations or between metastatic locations. Metastasis may be orig-
inated or evolved from a single clone or a cluster of subclones with
founder mutations differing from primary tumors, which can be
hardly removed by surgery. Metastatic tumor develop through the
linear or branched evolution which is more responsible for the
induction of drug resistance. Tumors with high instability of genes
evolve in the metastasis, of which a visible mass contains thou-
sands of cells resistant to drug [40]. The elimination of a subset
of metastatic lesions did not improve the long-term survival of the
patient, since the lesion commonly has 20 clonal genetic alterations
[41,42]. Mutations occurred in the founder cell of the metastasis,
which can be escaped from the primary tumor.
Cancer stem cells are a potential contributor for the heterogene-
ity and evolution of cell subpopulations or phenotypes [43,44]. The
heterogeneity among cancer stem cells in a tumor can be an another
driver to form the resistance by genetically distinct subclones
within the cancer stem cell compartment rather than in differenti-
ated compartments of a given tumor [45]. Single cell sequencing
can detect the heterogeneity between single cells [46,47]. Kim
et al. [48]examined the intratumoral heterogeneity between pri-
mary renal cell carcinoma and its lung metastasis using single-cell
RNA sequencing. At the single-cell level, metastatic renal carcinoma
cells showed an extensive heterogeneity in the activation of the
EGFR and Src pathways which needed a combinatorial regimen co-
targeting two mutually exclusive pathways for metastatic cancer
cells. The combinatorial strategy showed better efficacy than the
monotherapy during preclinical validation using patient-derived
xenograft in vitro and in vivo models.
5. Functional heterogeneity
The heterogeneity of gene-specific and associated functions,
e.g. epigenetic changes, posttranslational modifications, or cell dif-
ferentiation, exists and drives the initiation and progression of
cancer, where mutations occur more frequently than expected by
random chance [5,23,49]. Epigenetic molecular profiles of cancers
may reflect the variation of cell sub-compartments in normal cell
renewal systems from which cancers arise. Mutations in transcrip-
tional factors/regulators have tissue specificity, whereas histone
modifiers are often mutated across several cancer types [23,50].
Inherent functional variabilities in tumor propagation con-
tribute to cancer growth and therapy efficacy. Diffuse intrinsic
pontineglioma (DIPG) is a fatal brain cancer in the brainstem,
of which78% contained K27M mutation in H3F3A encoding his-
tone H3.3, or related HIST1H3B encoding histone H3.1 [51]. DIPG
H3-K27M subgroup was highly mutated in H3F3A, HIST1H3B, or
HIST1H3C and had highly unstable genomes [52]. DIPG has novel
oncogenic mutations connecting tumorigenesis, chromatin regu-
lation, or developmental signaling pathways. H3-K27M subgroup
may require multi-modal therapies due to the complexity and het-
erogeneity of the gene, if therapies targeting the histone mutations
become available.
Chromatin remodeling, histone and DNA methylation, acety-
lation, ubiquitination, or pre-mRNA splicing can be recurrently
16 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19
mutated in types of cancers [53,54]. The mutated pathways and
components include PI3K/AKT-mTOR signaling, KEAP1-NRF2-CUL3
apparatus, DNA methylation, p53-related pathways, or mRNA
processing. Correlations between DNA methylation with gene
mutation, expression and copy number profiles indicate clinical
risks for patients. Somatic alterations of 475 driver genes across
tumor genomes/exomes from nearly 7000 tumors were referred
for the drug ability and therapeutic design [55]. Driver gene muta-
tions may be subclonal in tumors and be benefited from targeted
therapy (Deciphering Anti-tumor Response with Intratumour Het-
erogeneity, NIH study trial registration number NCT02183883;
clintrials.gov). Targeting intratumor heterogeneity and temporal
acquisition of subclonal drivers and genomic cancer evolution
could impact therapeutic response, disease relapse, progression,
and development of precision medicine [56,57].
It is possible to identify driver mutations and causal determi-
nants of the disease, upstream genes of functional drivers, and
genetic determinants of tumor subtypes for designing new drugs
and detecting drug efficacy. Mutational intratumor heterogene-
ity was seen in tumor-suppressor genes, while SETD2, PTEN, and
KDM5C underwent multiple distinct and spatially separated inac-
tivating mutations within a single tumor [12]. Unfortunately, there
is few targeting drugs associated with suppressor genes to balance
driver mutations.
Regulatory networks are considered as an important approach
to understand the interaction between genes and proteins, define
signal pathways and factors involved in transcriptional regula-
tions and intracellular signals, and contribute to development and
promotion of tumor heterogeneity [58]. The variation of driver
factor distributions within the tumor attribute to the occurrence
of heterogeneity through driver factor signaling. Tumor growth
factor-␤-responding progenies differentiate and transcriptionally
activate p21, to stabilize NRF2 enhancing glutathione metabolism
and diminishing effectiveness of anti-cancer therapies [59]. It
can be a non-genetic paradigm that the driver factor signaling
contributes to tumor heterogeneity, characteristics, and drug resis-
tance.
6. Clinical potentials of precision medicine
Anticancer chemotherapeutics have severe side effects,
although having partial efficacy. Drugs targeting signaling onco-
proteins were expected to increase specificity and reduce side
effects, while easily develop the resistance [60]. The drug resistance
can be hardly predicted and protected, due to the limited under-
standing of signaling transduction, feedback loops, redundancy,
tumor heterogeneity, or target oncogenic roles. Structures of target
oncogenes and corresponding oncoproteins bound to the effectors
or regulators are critical for drug design. Therapeutic strategies
can be developed on basis of knowledge on precise mechanisms by
which target molecules can be activated or activate elements of the
downstream as a new way against cancers. One way to overcome
the drug resistance may need to define epigenetic alterations and
variations, through alterations of rapid kinetics, acquired drug
resistance, or genetic mutations. Transcriptional states through
reversible histone modifications and DNA methylation patterns
may cause a dynamic formation and development of tumor
heterogeneity.
The multidrug resistance transporter gene ABCB1 is one of
the most important mechanisms of drug resistance in cancer.
Histone acetyltransferases and histone deacetylases regulated
ABCB1 mRNA levels in human carcinoma cells [61]. The upreg-
ulation of ABCB1 by anticancer drugs was associated with a
dramatically induced acetyl-H3on the multidrug resistance1 P1
promoter [62,63]. Some transcription factors such as C/EBP␤
are directly involved in the development of acquired multidrug
resistance by trans-activating endogenous ABCB1 [64]. Mutation-
or epigenetics-associated or specific new drugs can target drug
resistance-associated/driven gene mutation or epigenomes to pre-
vent or reverse non-responsiveness to anti-cancer drugs. The
wild-type p53 represses ABCB1 promoter activity, while mutant
p53 acts oppositely. Preclinical evidence demonstrated that the
restoration or mimicry of wild-type p53 function could induce a
senescence program in liver carcinomas and sarcomas and apopto-
sis in lymphoma, which consequently leads to tumor clearance or
regression [65,66].
Another strategy is the combination of target therapies as a new
alternative for the treatment. The combination of ALK and MAPK
kinase inhibitors was active in an ALK-positive resistant tumor that
had developed a MAP2K1 activating mutation. The combination
of EGFR and fibroblast growth factor receptor (FGFR) inhibitors
was efficient in an EGFR mutant resistant cancer with a FGFR3
mutation [67]. Combined ALK and SRC (pp60c-src) inhibition was
effective in several ALK-driven diseases. Those preclinical evidence
demonstrates that the development of rational drug combinations
is critical in anti-cancer and anti-resistance strategies.
A new wave of small-molecular machineries interfere with
the functional heterogeneity of cancer cell synthetases [60].
Chemotherapeutics can target a broad spectrum of cellular
machineries (Fig. 3) by crosslinking nucleobases in DNA, block-
ing DNA replication, repair and nucleotide synthesis, interfering
with DNA polymerization, or typically using antagonists of ribonu-
cleotide reductases or thymidine synthetases. Drugs act on cellular
enzyme complexes and target chromatin modification, e.g. histone
deacetylases, bromodomain-containing proteins, or DNA methyl-
transferases (Fig. 2A). Drugs can inhibit HSP90, HSP70, HDAC6,
or different protease activities, e.g. chymotrypsin-like activity,
trypsin-like activity and/or caspase-like activity, within the 26S
proteasome to disrupt the protein degradation apparatus. The
ubiquitylation machinery and ubiquitin retrieval can also be
manipulated by small molecules, providing additional opportuni-
ties for interfering with proteasomal degradation (Fig. 2B).
Cancer cell signaling pathways and processes are important
components of functional heterogeneity. Mutated driver genes of
tumors were categorized into 12 pathways (Fig. 2C) to develop
agents and target functional heterogeneity of tumors [50]. Clinical
trials should be defined to match actionable oncogenic signatures
with personalized therapies. The interaction of genetic and non-
genetic determinants could influence the functional diversity and
therapy response for cancers in future.
7. Functional and precision biomarkers
There is a need to develop tumor heterogeneity-specific
biomarkers to detect the development and formation of the resis-
tance to target therapies, and drug-specific biomarkers to detect
drug efficacy, dynamics, metabolisms, pharmacology, or kinetics.
The epithelial-to-mesenchymal transition gene expression signa-
ture was used as a predictive biomarker of resistance to the EGFR
inhibitor erlotinib or PI3K/AKT–mTOR signaling inhibitors [68]. The
failure to assess mutations that drive resistance limits the ability
to predict response to treatment. Such limitations are also depen-
dent upon the sensitivity of the method or accuracy of mutations
present in less than 5% of cells. Chromosomal instability was con-
sidered as prognostic biomarker of solid tumor and may be related
to increased tumor cell heterogeneity to adapt to environmental
stresses [69]. We have emphasized the urgent need of biomarkers
to monitor the processes of tautomerism, depurination, deamina-
tion, slipped strand mispairing, error prone replication by-pass,
errors introduced during DNA repair, or induced mutation [70,71].
 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19 17

Fig. 2. The drug development strategies based on functional heterogeneity of tumors.

There are emerging evidences of functional heterogeneity of tumor, including DNA methylation, acetylation, network, or signal pathways. The functional heterogeneity could
potentially be drug development strategies. To address these heterogeneity through the following broad mechanisms: (A) for DNA replication and integrity,by crosslinking
nucleobases in DNA and blocking DNA replication; by inhibiting DNA repair; by interfering with the polymerization of DNA; and by inhibiting nucleotide synthesis, typically
using antagonists of ribonucleoside diphosphate reductaseor thymidine synthetase. Chromatin modification can be targeted by drugs to disrupt chromatin contacts or
affect the recruitment of nonhistone proteins to chromatin. The cellular enzyme complexes such as histone deacetylases (HDACs), bromodomain-containing proteins and
DNA methyltransferases are potential targets. (B) Protein chaperones assist in refolding mutated or stress-misfolded proteins. Complexes consist of the heat shock proteins
HSP90 and HSP70, as well as HSP70/HSP90 organizing protein (HOP), multipleco-chaperones, adaptor proteins, the ubiquitin E3 ligase CHIP(carboxy terminus of HSP70-
interacting protein) and the associated HDAC6 (Histone Deacetylase 6). Drugs can inhibit HSP90, HSP70 or HDAC6. Drugs can inhibit different protease activities – for example,
chymotrypsin-like activity, trypsin-like activity and/or caspase-like activity – within the 26S proteasome to disrupt the protein degradation apparatus. Ubiquitination can
affect proteins in many ways: it can signal for their degradation via the proteasome, alter their cellular location, affect their activity, and promote or prevent protein
interactions. The ubiquitylation machinery and ubiquitin retrieval can also be manipulated by small molecules, providing additional opportunities for interfering with
proteasomal degradation. Finally, the cancer cell signaling pathways and the cellular process they regulate compose the complexity of functional heterogeneity of tumors as
well. Generally,the mutated driver genes found in tumors can be classified into one or more of 12 pathways (C) including DNA damage control,chromatin modification, etc.
These pathways can be further organized into three core cellular processes (genome maintenance, cell fate and cell survival). Recognition of these pathways also allow the
development of agents that target functional heterogeneity of tumors.

There is an increasing evidence to screen therapeutic targets and
drug resistance-conferring gene mutations on circulating tumor
cells (CTC) and cell-free circulating tumor DNA (ctDNA) released
from metastatic deposits [72], although values and accuracy to
assess efficacy and resistance of new drugs or drug combinations
remain to be further validated. CTC or ctDNA presents in the major-
ity of patients with cancer, and has independent diagnostic and
prognostic significance for targeted therapies [73–75]. The “liquid
biopsy” brings challenges and opportunities related to intra-tumor
heterogeneity and recurrence of cancer, to identify genetic and
epigenetic aberrations and predict drug efficacy and acquired resis-
tance. The immune repertoire sequencing on T/B lymphocytes is
measured to determine B cell receptor and T cell receptor com-
plementary diversity decision region and assess the relationship
between the immune repertoire, drug efficacy, or disease [76,77].
Driver gene biomarkers can be developed by the integration of
genomic data on tumor heterogeneity, pinpoint driver genes from
functional RNAi screens, cell biology, clinical bioinformatics, drug
sensitivity and specificity, and prognosis. Clinical practices evi-
denced that the detection of gene mutations highly depended
upon the sensitivity of KRAS measurements in EGFR-non-sensitive
patients with metastatic colorectal cancer [78,79].
8. Opportunities and challenges of precision therapy
The high-throughput sequencing revealed that several hundred
gene mutations underlie tumor heterogeneity. Those findings high-
light the difficulties in developing newer therapies for the disease
and suggest that targeted approaches will rely on better under-
standing of patient’s tumor. One of the resistance causes to targeted
therapies can be a challenge or invisible obstacle for achievement
of precision medicine or “Moonshot” plan . Fig. 3 furthermore
describes the influence of tumor heterogeneity in drug efficacy to
bring opportunities and challenges for precision therapy.
Tumor heterogeneity should be closely tracked to identify trunk
or heterogeneous events by more precise methodologies for under-
standing of resistance mechanisms. The heterogeneity should be
qualified for optimal designs of combination therapies to counter-
act the sub-modal resistant mutation. Specific and efficient drugs
should be discovered and validated to target the potential escape
mechanisms and non-drugable driver mutations. Precise biomark-
ers should be developed to monitor precision therapy and subclonal
dynamics of tumor architecture. Thus, we believe the heterogene-
ity can be a breakthrough point to perform precision medicine to
cure patients with cancer.
18 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19

Fig. 3. The challenges and opportunities presented by tumor heterogeneity for precision medicine.
Mutations introduced during primary tumor cell growth result in clonal and metastatic heterogeneity. At the right part, PI3K as representative signal transduction pathways
affected by mutations is illustrated. Activated PI3K-Akt signaling modulates the function of numerous substrates involved in the regulation of cellular growth. PI3K/Akt
signalling pathway components are frequently altered in human cancers. Here we take PI3K as an example to discuss the challenges and opportunities presented by tumor
heterogeneity for precision medicine as following: (1) Tracking Heterogeneity: Biopsies in one region of a heterogeneous tumor will identify trunk events but may also
identify more heterogeneous events. Comparison of paired metastatic samples may enhance the identification of trunk events for therapeutic targeting. (2) Quantifying
Heterogeneity: Measureing the extent ot heterogeneity in cancers will probably lead to rational design of combination therapies to counteract the sub-modal resistant
mutaiton. (3) Clarifying resistance mechanisms: A cataloguing of the potential resistance mechanisms-genetic,epigenetic and adaptive-to specific targeted therapies for each
individual oncogenic aberration. (4) Developing effective agents: The development of a panel of therapeutic agents to target the potential escape mechanisms, and to target
what are at present regarded as non-drugable driver mutations. (5) Monitoring drug efficacy: Developing of precision biomarkers to measure precision medicine and monitor
the subclonal dynamics of tumor architecture.

Conflict of interest
The authors declare that there are no conflicts of interest.
Acknowledgements
The work was supported by Shanghai Leading Academic
Discipline Project (B115), Zhongshan Distinguished Professor
Grant (XDW), The National Nature Science Foundation of
China (91230204, 81270099, 81320108001, 81270131, 81400035,
81570075, 81500058, 81500025), The Shanghai Committee of Sci-
ence and Technology (12JC1402200, 12431900207, 11410708600),
Zhejiang Provincial Natural Science Foundation (Z15H010002),
Zhejiang Provincial Science Technology Department Foundation
(WKJ-ZJ-1526).
References
[1] X.D. Wang, Gene mutation-based and specific therapies in precision
medicine, J. Cell Mol. Med. 20 (April ()4) (2016) 577–580, http://dx.doi.org/10.
1111/jcmm.12722.
[2] F.S. Collins, H. Varmus, A new initiative on precision medicine, N. Engl. J. Med.
372 (9) (2015) 793–795.
[3] E.A. Ashley, The precision medicine initiative: a new national effort, JAMA 313
(21) (2015) 2119–2120.
[4] C. Chen, et al., Five critical elements to ensure the precision medicine, Cancer
Metastasis Rev. 34 (2) (2015) 313–318.
[5] M.S. Lawrence, et al., Mutational heterogeneity in cancer and the search for
new cancer-associated genes, Nature 499 (7457) (2013) 214–218.
[6] C. Swanton, Intratumor heterogeneity: evolution through space and time,
Cancer Res. 72 (19) (2012) 4875–4882.
[7] M. Dolsten, M. Søgaard, Precision medicine: an approach to R&D for
delivering superior medicines to patients, Clin. Transl. Med. 1 (1) (2012) 7,
http://dx.doi.org/10.1186/2001-1326-1-7.
[8] M. He, J. Xia, M. Shehab, X.D. Wang, The development of precision medicine in
clinical practice, Clin. Transl. Med. 4 (December (1)) (2015) 69, http://dx.doi.
org/10.1186/s40169-015-0069-y.
[9] C. Holohan, et al., Cancer drug resistance: an evolving paradigm, Nat. Rev
Cancer 13 (10) (2013) 714–726.
[10] N.C. Turner, J.S. Reis-Filho, Genetic heterogeneity and cancer drug resistance,
Lancet Oncol. 13 (4) (2012) e178–85.
[11] F. Niu, D.C. Wang, J.P. Lu, W. Wu, X.D. Wang, Potentials of single-cell biology
in identification and validation of disease biomarkers, J. Cell. Mol. Med. (April)
(2016), http://dx.doi.org/10.1111/jcmm.12868.
[12] P.D. Thomas, M. Kahn, Kat3 coactivators in somatic stem cells and cancer
stem cells: biological roles, evolution, and pharmacologic manipulation, Cell
Biol. Toxicol. 32 (February (1)) (2016) 61–81, http://dx.doi.org/10.1007/
s10565-016-9318-0.
[13] N. McGranahan, C. Swanton, Biological and therapeutic impact of intratumor
heterogeneity in cancer evolution, Cancer Cell 27 (1) (2015) 15–26.
[14] X.D. Wang, Cancer Moonshot 2020: a new march of clinical and translational
medicine, Clin. Transl. Med. 5 (2016) 11.
[15] Y. Zhu, M. Luo, M. Brooks, S.G. Clouthier, M.S. Wicha, Biological and clinical
significance of cancer stem cell plasticity, Clin. Transl. Med. 3 (December (1))
(2014) 32, http://dx.doi.org/10.1186/s40169-014-0032-3.
[16] X.D. Wang, K. Li, H. Chen, D. Wang, Y. Zhang, C. Bai, Does hepatocyte growth
factor/c-Met signal play synergetic role in lung cancer? J. Cell. Mol. Med. 14
(4) (2010) 833–839.
[17] M. Zwitter, M. Rajer, K. Stanic, et al., Intercalated chemotherapy and erlotinib
for non-small cell lung cancer (NSCLC) with activating epidermal growth
factor receptor (EGFR) mutations, Cancer Biol. Ther. (June) (2016) 0.
[18] R. Liu, X.D. Wang, K. Aihara, L. Chen, Early diagnosis of complex diseases by
molecular biomarkers: network biomarkers, and dynamical network
biomarkers, Med. Res. Rev. 34 (3) (2014) 455–478.
 D. Wu et al. / Seminars in Cancer Biology 42 (2017) 13–19
[19] C.A. Klein, N.H. Stoecklein, Lessons from an aggressive cancer: evolutionary
dynamics in esophageal carcinoma, Cancer Res. 69 (13) (2009) 5285–5288.
[20] X. Wu, L. Chen, X.D. Wang, Network biomarkers, interaction networks and
dynamical network biomarkers in respiratory diseases, Clin. Transl. Med.
(June (3)) (2014) 16.
[21] L.A. Loeb, Human cancers express mutator phenotypes: origin, consequences
and targeting, Nat. Rev. Cancer 11 (6) (2011) 450–457.
[22] S. Devarakonda, D. Morgensztern, R. Govindan, Genomic alterations in lung
adenocarcinoma, Lancet Oncol. 16 (7) (2015) e342–51.
[23] C. Kandoth, et al., Mutational landscape and significance across 12 major
cancer types, Nature 502 (7471) (2013) 333–339.
[24] P.J. Stephens, et al., The landscape of cancer genes and mutational processes
in breast cancer, Nature 486 (7403) (2012) 400–404.
[25] T. Shimizu, et al., The clinical effect of the dual-targeting strategy involving
PI3K/AKT/mTOR and RAS/MEK/ERK pathways in patients with advanced
cancer, Clin. Cancer Res. 18 (8) (2012) 2316–2325.
[26] E. Hodis, et al., A landscape of driver mutations in melanoma, Cell 150 (2)
(2012) 251–263.
[27] M.B. Carper, P.P. Claudio, Clinical potential of gene mutations in lung cancer,
Clin. Transl. Med. 4 (1) (2015) 33, http://dx.doi.org/10.1186/s40169-015-
0074-1.
[28] A. Kreso, et al., Variable clonal repopulation dynamics influence
chemotherapy response in colorectal cancer, Science 339 (6119) (2013)
543–548.
[29] T. Zeng, D.C. Wang, X.D. Wang, F. Xu, L. Chen, Prediction of dynamical drug
sensitivity and resistance by module network rewiring-analysis based on
transcriptional profiling, Drug Resist. Updat. 17 (3) (2014 Jul) 64–76.
[30] M.B. Meads, R.A. Gatenby, W.S. Dalton, Environment-mediated drug
resistance: a major contributor to minimal residual disease, Nat. Rev. Cancer 9
(9) (2009) 665–674.
[31] R. Straussman, et al., Tumour micro-environment elicits innate resistance to
RAF inhibitors through HGF secretion, Nature 487 (7408) (2012) 500–504.
[32] T.R. Wilson, et al., Widespread potential for growth-factor-driven resistance
to anticancer kinase inhibitors, Nature 487 (7408) (2012) 505–509.
[33] L. Boussemart, et al., eIF4F is a nexus of resistance to anti-BRAF and anti-MEK
cancer therapies, Nature 513 (7516) (2014) 105–109.
[34] M.R. Junttila, F.J. de Sauvage, Influence of tumour micro-environment
heterogeneity on therapeutic response, Nature 501 (7467) (2013)
346–354.
[35] U.M. Polanska, A. Orimo, Carcinoma-associated fibroblasts: non-neoplastic
tumour-promoting mesenchymal cells, J. Cell. Physiol. 228 (8) (2013)
1651–1657.
[36] R. Berardi, F. Morgese, A. Onofri, et al., Role of maspin in cancer, Clin. Transl.
Med. 2 (1) (2013) 8, http://dx.doi.org/10.1186/2001-1326-2-8.
[37] P.S. Hegde, et al., Predictive impact of circulating vascular endothelial growth
factor in four phase III trials evaluating bevacizumab, Clin. Cancer Res. 19 (4)
(2013) 929–937.
[38] J. Zhou, Z.T. Zhu, C.X. Ba, H.Z. Sun, X.D. Wang, Proteomic profiling of
lymphocytes in autoimmunity, inflammation and cancer, J. Transl. Med. 12
(2014) 6.
[39] D.M. Pardoll, The blockade of immune checkpoints in cancer immunotherapy,
Nat. Rev. Cancer 12 (4) (2012) 252–264.
[40] B.J. Monk, T.J. Herzog, K.S. Tewari, Evolution of Chemosensitivity and
resistance assays as predictors of clinical outcomes in epithelial ovarian
cancer patients, Curr. Pharm. Des. (May) (2016) 41.
[41] S. Heerboth, G. Housman, M. Leary, et al., EMT and tumor metastasis, Clin.
Transl. Med. 4 (2015) 6, http://dx.doi.org/10.1186/s40169-015-0048-3.
[42] M.F. Clarke, et al., Cancer stem cells—perspectives on current status and
future directions: AACR Workshop on cancer stem cells, Cancer Res. 66 (19)
(2006) 9339–9344.
[43] J. Gu, X.D. Wang, New future of cell biology and toxicology: thinking deeper,
Cell Biol. Toxicol. 1 (2016) 1–3, http://dx.doi.org/10.1007/s10565-016-9313-5
(Epub 2016 Feb 13).
[44] K. Anderson, et al., Genetic variegation of clonal architecture and propagating
cells in leukaemia, Nature 469 (7330) (2011) 356–361.
[45] B. Bakker, A. Taudt, M.E. Belderbos, et al., Single-cell sequencing reveals
karyotype heterogeneity in murine and human malignancies, Genome Biol.
17 (1) (2016) 115, http://dx.doi.org/10.1186/s13059-016-0971-7.
[46] S. Ling, et al., Extremely high genetic diversity in a single tumor points to
prevalence of non-Darwinian cell evolution, Proc. Natl. Acad. Sci. U. S. A. 112
(47) (2015) E6496–E6505.
[47] K.T. Kim, et al., Application of single-cell RNA sequencing in optimizing a
combinatorial therapeutic strategy in metastatic renal cell carcinoma,
Genome Biol. 17 (1) (2016) 80.
[49] B.L. Brücher, Y. Li, P. Schnabel, et al., Genomics, microRNA, epigenetics, and
proteomics for future diagnosis, treatment and monitoring response in upper
GI cancers, Clin. Transl. Med. 5 (1) (2016) 13, http://dx.doi.org/10.1186/
s40169-016-0093-6.
19
[50] B. Vogelstein, et al., Cancer genome landscapes, Science 339 (6127) (2013)
1546–1558.
[51] G. Wu, et al., Somatic histone H3 alterations in pediatric diffuse intrinsic
pontine gliomas and non-brainstem glioblastomas, Nat. Genet. 44 (3) (2012)
251–253.
[52] P. Buczkowicz, et al., Genomic analysis of diffuse intrinsic pontine gliomas
identifies three molecular subgroups and recurrent activating ACVR1
mutations, Nat. Genet. 46 (5) (2014) 451–456.
[53] I.R. Watson, et al., Emerging patterns of somatic mutations in cancer, Nat. Rev.
Genet. 14 (10) (2013) 703–718.
[54] Y. Sato, et al., Integrated molecular analysis of clear-cell renal cell carcinoma,
Nat. Genet. 45 (8) (2013) 860–867.
[55] C. Rubio-Perez, et al., In silico prescription of anticancer drugs to cohorts of 28
tumor types reveals targeting opportunities, Cancer Cell 27 (3) (2015)
382–396.
[56] C. Hiley, et al., Deciphering intratumor heterogeneity and temporal
acquisition of driver events to refine precision medicine, Genome Biol. 15 (8)
(2014) 453.
[57] M.et al. Jamal-Hanjani, Tracking genomic cancer evolution for precision
medicine: the lung TRACERx study, PLoS Biol. 12 (7) (2014) e1001906.
[58] X. Liu, J. Wang, L. Chen, Whole-exome sequencing reveals recurrent somatic
mutation networks in cancer, Cancer Lett. 340 (2) (2013) 270–276.
[59] N. Oshimori, D. Oristian, E. Fuchs, TGF-beta promotes heterogeneity and drug
resistance in squamous cell carcinoma, Cell 160 (5) (2015) 963–976.
[60] M. Dobbelstein, U. Moll, Targeting tumour-supportive cellular machineries in
anticancer drug development, Nat. Rev. Drug Discov. 13 (3) (2014) 179–196.
[61] S. Jin, K.W. Scotto, Transcriptional regulation of the MDR1 gene by histone
acetyltransferase and deacetylase is mediated by NF-Y, Mol. Cell. Biol. 18 (7)
(1998) 4377–4384.
[62] Y.H. Kim, D. Jeoung, Tubulin Beta3 serves as a target of HDAC3 and mediates
resistance to microtubule-targeting drugs, Mol. Cells 38 (8) (2015) 705–714,
http://dx.doi.org/10.14348/molcells.2015.0086.
[63] E.K. Baker, et al., Epigenetic changes to the MDR1 locus in response to
chemotherapeutic drugs, Oncogene 24 (54) (2005) 8061–8075.
[64] G.K. Chen, et al., CCAAT/enhancer-binding protein beta (nuclear factor for
interleukin 6) transactivates the human MDR1 gene by interaction with an
inverted CCAAT box in human cancer cells, Mol. Pharmacol. 65 (4) (2004)
906–916.
[65] W. Xue, et al., Senescence and tumour clearance is triggered by p53
restoration in murine liver carcinomas, Nature 445 (7128) (2007) 656–660.
[66] A. Ventura, et al., Restoration of p53 function leads to tumour regression in
vivo, Nature 445 (7128) (2007) 661–665.
[67] A.S. Crystal, et al., Patient-derived models of acquired resistance can identify
effective drug combinations for cancer, Science 346 (6216) (2014) 1480–1486.
[68] L.A. Byers, et al., An epithelial-mesenchymal transition gene signature
predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a
therapeutic target for overcoming EGFR inhibitor resistance, Clin. Cancer Res.
19 (1) (2013) 279–290.
[69] C. Swanton, et al., Chromosomal instability determines taxane response, Proc.
Natl. Acad. Sci. U. S. A. 106 (21) (2009) 8671–8676.
[70] D. Wu, X. Wang, Application of clinical bioinformatics in lung cancer-specific
biomarkers, Cancer Metastasis Rev. 34 (2) (2015) 209–216.
[71] X. Wang, A.A. Adjei, Lung cancer and metastasis: new opportunities and
challenges, Cancer Metastasis Rev. 34 (2) (2015) 169–171.
[72] K. Pantel, C. Alix-Panabieres, Real-time liquid biopsy in cancer patients: fact
or fiction? Cancer Res. 73 (21) (2013) 6384–6388.
[73] M. Westphal, K. Lamszus, Circulating biomarkers for gliomas, Nat. Rev.
Neurol. 11 (10) (2015) 556–566.
[74] J. Tabernero, et al., Analysis of circulating DNA and protein biomarkers to
predict the clinical activity of regorafenib and assess prognosis in patients
with metastatic colorectal cancer: a retrospective, exploratory analysis of the
CORRECT trial, Lancet Oncol. 16 (8) (2015) 937–948.
[75] P.P. Lin, Integrated EpCAM-independent subtraction enrichment and iFISH
strategies to detect and classify disseminated and circulating tumors cells,
Clin. Transl. Med. 4 (1) (2015) 38, http://dx.doi.org/10.1186/s40169-015-
0081-2.
[76] J.R. McDaniel, et al., Ultra-high-throughput sequencing of the immune
receptor repertoire from millions of lymphocytes, Nat. Protoc. 11 (3) (2016)
429–442.
[77] H. Robins, Immunosequencing: applications of immune repertoire deep
sequencing, Curr. Opin. Immunol. 25 (5) (2013) 646–652.
[78] F. Molinari, et al., Increased detection sensitivity for KRAS mutations
enhances the prediction of anti-EGFR monoclonal antibody resistance in
metastatic colorectal cancer, Clin. Cancer Res. 17 (14) (2011) 4901–4914.
[79] W. De Roock, et al., Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the
efficacy of cetuximab plus chemotherapy in chemotherapy-refractory
metastatic colorectal cancer: a retrospective consortium analysis, Lancet
Oncol. 11 (8) (2010) 753–762.