Distinct subclonal tumour responses to therapy revealed by circulating cell-free DNA.

BACKGROUND: The application of precision medicine in oncology requires in-depth characterisation of a patient's tumours and the dynamics of their responses to treatment. PATIENTS AND METHODS: We used next-generation sequencing of circulating cell-free DNA (cfDNA) to monitor the response of a KIT p.L576P-mutant metastatic vaginal mucosal melanoma to sequential targeted, immuno- and chemotherapy. RESULTS: Despite a KIT mutation, the response to imatinib was mixed. Unfortunately, tumours were not accessible for molecular analysis. To study the mechanism underlying the mixed clinical response, we carried out whole-exome sequencing and targeted longitudinal analysis of cfDNA. This revealed two tumour subclones; one with a KIT mutation that responded to imatinib and a second KIT-wild-type subclone that did not respond to imatinib. Notably, the subclones also responded differently to immunotherapy. However, both subclones responded to carboplatin/paclitaxel, and although the KIT-wild-type subclone progressed after chemotherapy, it responded to subsequent re-administration of paclitaxel. CONCLUSION: We show that cfDNA can reveal tumour evolution and subclonal responses to therapy even when biopsies are not available.

Tumourigenic non-small-cell lung cancer mesenchymal circulating tumour cells: a clinical case study.

BACKGROUND: Over the past decade, numerous reports describe the generation and increasing utility of non-small-cell lung cancer (NSCLC) patient-derived xenografts (PDX) from tissue biopsies. While PDX have proven useful for genetic profiling and preclinical drug testing, the requirement of a tissue biopsy limits the available patient population, particularly those with advanced oligometastatic disease. Conversely, 'liquid biopsies' such as circulating tumour cells (CTCs) are minimally invasive and easier to obtain. Here, we present a clinical case study of a NSCLC patient with advanced metastatic disease, a never smoker whose primary tumour was EGFR and ALK wild-type. We demonstrate for the first time, tumorigenicity of their CTCs to generate a patient CTC-derived eXplant (CDX). PATIENTS AND METHODS: CTCs were enriched at diagnosis and again 2 months later during disease progression from 10 ml blood from a 48-year-old NSCLC patient and implanted into immunocompromised mice. Resultant tumours were morphologically, immunohistochemically, and genetically compared with the donor patient's diagnostic specimen. Mice were treated with cisplatin and pemetrexed to assess preclinical efficacy of the chemotherapy regimen given to the donor patient. RESULTS: The NSCLC CDX expressed lung lineage markers TTF1 and CK7 and was unresponsive to cisplatin and pemetrexed. Examination of blood samples matched to that used for CDX generation revealed absence of CTCs using the CellSearch EpCAM-dependent platform, whereas size-based CTC enrichment revealed abundant heterogeneous CTCs of which ∼80% were mesenchymal marker vimentin positive. Molecular analysis of the CDX, mesenchymal and epithelial CTCs revealed a common somatic mutation confirming tumour origin and showed CDX RNA and protein profiles consistent with the predominantly mesenchymal phenotype. CONCLUSIONS: This study shows that the absence of NSCLC CTCs detected by CellSearch (EpCAM(+)) does not preclude CDX generation, highlighting epithelial to mesenchymal transition and the functional importance of mesenchymal CTCs in dissemination of this disease.

Ligation of the CD2 co-stimulatory receptor enhances IL-2 production from first-generation chimeric antigen receptor T cells.

T cells bearing chimeric antigen receptors (CARs) are broadly categorised into first- and second-generation receptors. Second-generation CARs contain a co-stimulatory signalling molecule and have been shown to secrete IL-2, undergo greater proliferation and to have enhanced persistence in vivo. However, we have shown that T cells bearing a first-generation CAR containing a CD19-targeting scFv (single-chain variable fragment) and the CD3ζ-signalling domain are able to produce IL-2 upon co-culture with CD19(+) B-cell lymphomas independent of CD28 activity. Here, we report that signalling through endogenous CD2 following ligation with its ligands, CD48 in mouse and CD58 in humans, drives IL-2 production by first-generation CD19-specific CAR. Moreover, the high levels of IL-2 produced by human T cells engrafted with a second-generation CD28-containing CAR during target-cell recognition are dependent to a degree upon CD2 receptor activity. These observations highlight the fact that the functional activity induced by T-cell-expressed CARs is dependent upon endogenous 'natural' receptor interactions. A deeper understanding of the role of these activities will serve to further refine the design of future CARs to either exploit or avoid these interactions.

Substitution of Asp-210 in HAP1 (APE/Ref-1) eliminates endonuclease activity but stabilises substrate binding.

HAP1, also known as APE/Ref-1, is the major apurinic/apyrimidinic (AP) endonuclease in human cells. Previous structural studies have suggested a possible role for the Asp-210 residue of HAP1 in the enzymatic function of this enzyme. Here, we demonstrate that substitution of Asp-210 by Asn or Ala eliminates the AP endonuclease activity of HAP1, while substitution by Glu reduces specific activity approximately 500-fold. Nevertheless, these mutant proteins still bind efficiently to oligonucleotides containing either AP sites or the chemically unrelated bulky p-benzoquinone (pBQ) derivatives of dC, dA and dG, all of which are substrates for HAP1. These results indicate that Asp-210 is required for catalysis, but not substrate recognition, consistent with enzyme kinetic data indicating that the HAP1-D210E protein has a 3000-fold reduced K(cat )for AP site cleavage, but an unchanged K(m). Through analysis of the binding of Asp-210 substitution mutants to oligonucleotides containing either an AP site or a pBQ adduct, we conclude that the absence of Asp-210 allows the formation of a stable HAP1-substrate complex that exists only transiently during the catalytic cycle of wild-type HAP1 protein. We interpret these data in the context of the structure of the HAP1 active site and the recently determined co-crystal structure of HAP1 bound to DNA substrates.

The structure and functions of the HAP1/Ref-1 protein.

The HAP1/Ref-1 (hereafter referred to as HAP1) protein is a nuclear enzyme that apparently performs two distinct roles in the cellular defense against oxidative stress. One well-established role is in the repair of a variety of lesions induced in DNA either by spontaneous hydrolysis or by reactive oxygen species (ROS). This function has been characterized in great detail and the roles played by individual active site amino acid residues have been defined. The second role, which was identified only relatively recently and is still not characterized in detail, is to regulate the DNA binding activity of a group of nuclear factors. This second role proceeds via the modification of the oxidation/reduction (redox) state of a cysteine residue in the target protein, although the mechanism by which this is achieved remains to be elucidated. In this article, we shall review the latest knowledge on the relationship between structure and the dual functions of HAP1, and we will seek to explain in detail the roles played by several individual amino acid residues in the catalytic function of the HAP1 protein.

Asparagine 212 is essential for abasic site recognition by the human DNA repair endonuclease HAP1.

HAP1 is a divalent cation-dependent endonuclease from human cells with specificity for apurinic/apyrimidinic (AP) sites in DNA. Extraction of the essential metal ion from purified HAP1 stabilized its binding to an oligonucleotide containing a single AP site, permitting AP site binding studies to be undertaken using gel retardation assays. Binding of HAP1 to such an oligonucleotide was dependent upon the presence of an AP site. Previous structural and modelling studies have suggested a role for Asn212 (Asn153 in exonuclease III, the bacterial homologue of HAP1) in substrate recognition. Substitution of alanine for Asn212 abolished the AP endonuclease activity of purified recombinant HAP1 protein. More conservative substitutions of aspartate or glutamine for Asn212 still led to a reduction in specific activity of at least 300-fold. Moreover, none of the three Asn212 substitution mutants of HAP1 possessed detectable AP site binding activity in vitro. This study indicates that chelation of the active site metal ion in HAP1 stabilizes the complex of the protein with AP sites and identifies an active site asparagine residue as an important component of AP site recognition by the HAP1 protein.

Evidence for a common active site for cleavage of an AP site and the benzene-derived exocyclic adduct, 3,N4-benzetheno-dC, in the major human AP endonuclease.

We have previously reported that the 3,N4-benzetheno-dC (p-BQ-dC) endonuclease activity found in HeLa cells is a novel function of the major human AP endonuclease (HAP1) [Hang et al. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 13737-13741]. In this study, we compare the enzymatic and biochemical properties of the enzyme toward p-BQ-dC and an AP site in a defined oligonucleotide. A comparative analysis of the specificity constants (Kcat./Km) for p-BQ-dC and an AP site indicates that the AP site is the preferred substrate. The enzyme does not cleave other structurally related exocyclic adducts and modified nucleosides such as 1,N6-etheno-dA, 3,N4-etheno-dC, 1, N2-etheno-dG, 1,N2-propano-dG, 8-oxo-dG, and thymine glycol. The p-BQ-dC activity requires a double-stranded DNA substrate and is affected by the base in the opposite strand, with maximal activity for a p-BQ-dC.G pair and minimal activity for a p-BQ-dC.C pair. The p-BQ-dC activity also requires Mg2+, Mn2+, or Zn2+ with optimal concentration spectra similar to those for the AP function. The optimal pH ranges for these two functions are also similar to each other (5.5-6.5). Six mutant HAP1 proteins containing single amino acid substitutions were assayed in parallel for comparison of their activities toward p-BQ-dC and the AP site. These mutants either concomitantly lost (N212A, D210N) or had reduced (D219A, E96A, and N212Q) or unchanged (H116N) p-BQ-dC and AP activities. This parallelism strongly supports the hypothesis that cleavage of p-BQ-dC requires the same catalytic active site as that proposed for the AP function. This dual activity toward two structurally unrelated substrates, an AP site and a bulky exocyclic adduct, has implications for substrate recognition. The AP site and p-BQ-dC cause different changes in the local conformation around the lesion as it is visualized by molecular modeling.

Detection of N-Ras codon 61 mutations in subpopulations of tumor cells in multiple myeloma at presentation.

Activating point mutations in codons 12, 13, or 61 of the K-ras and N-ras genes have been reported to occur in up to 40% of patients with multiple myeloma at presentation. In a study of 34 presentation myeloma cases using a sensitive polymerase chain reaction-restriction fragment length polymorphism strategy on enriched tumor cell populations, the present study detected N-ras codon 61 mutation-positive cells in all patients. Quantitative plaque hybridization using allele-specific oligonucleotide probes showed that in the majority of patients, ras mutation-positive cells comprise only a subpopulation of the total malignant plasma cell compartment (range, 12%-100%). Using clonospecific point mutations in the 5' untranslated region of the BCL6 gene to quantitate clonal B cells in FACS-sorted bone marrow populations from 2 patients, the representation of ras mutation-positive cells was independent of immunophenotype. These observations imply that mutational activation of N-ras codon 61 is a mandatory event in the pathogenesis of multiple myeloma; such mutations provide a marker of intraclonal heterogeneity that may originate at an earlier ontologic stage than immunophenotypic diversification of the malignant B cell clone.

The crystal structure of the human DNA repair endonuclease HAP1 suggests the recognition of extra-helical deoxyribose at DNA abasic sites.

The structure of the major human apurinic/ apyrimidinic endonuclease (HAP1) has been solved at 2.2 A resolution. The enzyme consists of two symmetrically related domains of similar topology and has significant structural similarity to both bovine DNase I and its Escherichia coli homologue exonuclease III (EXOIII). A structural comparison of these enzymes reveals three loop regions specific to HAP1 and EXOIII. These loop regions apparently act in DNA abasic site (AP) recognition and cleavage since DNase I, which lacks these loops, correspondingly lacks AP site specificity. The HAP1 structure furthermore suggests a mechanism for AP site binding which involves the recognition of the deoxyribose moiety in an extrahelical conformation, rather than a 'flipped-out' base opposite the AP site.

Cloning and characterization of a functional human homolog of Escherichia coli endonuclease III.

Repair of oxidative damage to DNA bases is essential to prevent mutations and cell death. Endonuclease III is the major DNA glycosylase activity in Escherichia coli that catalyzes the excision of pyrimidines damaged by ring opening or ring saturation, and it also possesses an associated lyase activity that incises the DNA backbone adjacent to apurinic/apyrimidinic sites. During analysis of the area adjacent to the human tuberous sclerosis gene (TSC2) in chromosome region 16p13.3, we identified a gene, OCTS3, that encodes a 1-kb transcript. Analysis of OCTS3 cDNA clones revealed an open reading frame encoding a predicted protein of 34.3 kDa that shares extensive sequence similarity with E. coli endonuclease III and a related enzyme from Schizosaccharomyces pombe, including a conserved active site region and an iron/sulfur domain. The product of the OCTS3 gene was therefore designated hNTH1 (human endonuclease III homolog 1). The hNTH1 protein was overexpressed in E. coli and purified to apparent homogeneity. The recombinant protein had spectral properties indicative of the presence of an iron/sulfur cluster, and exhibited DNA glycosylase activity on double-stranded polydeoxyribonucleotides containing urea and thymine glycol residues, as well as an apurinic/apyrimidinic lyase activity. Our data indicate that hNTH1 is a structural and functional homolog of E. coli endonuclease III, and that this class of enzymes, for repair of oxidatively damaged pyrimidines in DNA, is highly conserved in evolution from microorganisms to human cells.