Plasma cell-free DNA hydroxymethylation profiling reveals anti-PD-1 treatment response and resistance biology in non-small cell lung cancer

BACKGROUND

Treatment with immune checkpoint inhibitors (ICIs) targeting programmed death-1 (PD-1) can yield durable antitumor responses, yet not all patients respond to ICIs. Current approaches to select patients who may benefit from anti-PD-1 treatment are insufficient. 5-hydroxymethylation (5hmC) analysis of plasma-derived cell-free DNA (cfDNA) presents a novel non-invasive approach for identification of therapy response biomarkers which can tackle challenges associated with tumor biopsies such as tumor heterogeneity and serial sample collection.

METHODS

151 blood samples were collected from 31 patients with non-small cell lung cancer (NSCLC) before therapy started and at multiple time points while on therapy. Blood samples were processed to obtain plasma-derived cfDNA, followed by enrichment of 5hmC-containing cfDNA fragments through biotinylation via a two-step chemistry and binding to streptavidin coated beads. 5hmC-enriched cfDNA and whole genome libraries were prepared in parallel and sequenced to obtain whole hydroxymethylome and whole genome plasma profiles, respectively.

RESULTS

Comparison of on-treatment time point to matched pretreatment samples from same patients revealed that anti-PD-1 treatment induced distinct changes in plasma cfDNA 5hmC profiles of responding patients, as judged by Response evaluation criteria in solid tumors, relative to non-responders. In responders, 5hmC accumulated over genes involved in immune activation such as inteferon (IFN)-γ and IFN-α response, inflammatory response and tumor necrosis factor (TNF)-α signaling, whereas in non-responders 5hmC increased over epithelial to mesenchymal transition genes. Molecular response to anti-PD-1 treatment, as measured by 5hmC changes in plasma cfDNA profiles were observed early on, starting with the first cycle of treatment. Comparison of pretreatment plasma samples revealed that anti-PD-1 treatment response and resistance associated genes can be captured by 5hmC profiling of plasma-derived cfDNA. Furthermore, 5hmC profiling of pretreatment plasma samples was able to distinguish responders from non-responders using T cell-inflamed gene expression profile, which was previously identified by tissue RNA analysis.

CONCLUSIONS

These results demonstrate that 5hmC profiling can identify response and resistance associated biological pathways in plasma-derived cfDNA, offering a novel approach for non-invasive prediction and monitoring of immunotherapy response in NSCLC.

Epigenomic Blood-Based Early Detection of Pancreatic Cancer Employing Cell-Free DNA

BACKGROUND & AIMS

Early detection of pancreatic cancer (PaC) can drastically improve survival rates. Approximately 25% of subjects with PaC have type 2 diabetes diagnosed within 3 years prior to the PaC diagnosis, suggesting that subjects with type 2 diabetes are at high risk of occult PaC. We have developed an early-detection PaC test, based on changes in 5-hydroxymethylcytosine (5hmC) signals in cell-free DNA from plasma.

METHODS

Blood was collected from 132 subjects with PaC and 528 noncancer subjects to generate epigenomic and genomic feature sets yielding a predictive PaC signal algorithm. The algorithm was validated in a blinded cohort composed of 102 subjects with PaC, 2048 noncancer subjects, and 1524 subjects with non-PaCs.

RESULTS

5hmC differential profiling and additional genomic features enabled the development of a machine learning algorithm capable of distinguishing subjects with PaC from noncancer subjects with high specificity and sensitivity. The algorithm was validated with a sensitivity for early-stage (stage I/II) PaC of 68.3% (95% confidence interval [CI], 51.9%-81.9%) and an overall specificity of 96.9% (95% CI, 96.1%-97.7%).

CONCLUSIONS

The PaC detection test showed robust early-stage detection of PaC signal in the studied cohorts with varying type 2 diabetes status. This assay merits further clinical validation for the early detection of PaC in high-risk individuals.

Early detection of pancreatic cancer using 5-hydroxymethylation profiles in plasma-derived cell-free DNA

672

Background: Pancreatic cancer is one of the deadliest cancers, with approximately 15-20% of patients who present at diagnosis with a resectable disease. The major barrier to better outcomes is the lack of early-detection molecular tools to enable timely intervention. We have developed a test that enables the detection of pancreatic cancer from a simple blood draw. The test incorporates a novel, genome-wide sequencing-based epigenomics detection method that enriches for DNA loci that undergo active de-methylation. The measurement of 5-hydroxymethylcytosine (5hmC) provides a unique and stable biomarker for the early detection of cancer including pancreatic cancer. Methods: Whole-blood was obtained from a training cohort of 660 individuals (consisting of 132 pancreatic cancers (PaCa) and 528 non-cancers) and a validation cohort of 2,150 individuals (consisting of 102 PaCa and 2,048 non-cancers). Cell-free DNA (cfDNA) was isolated from plasma from which 5hmC and whole-genome libraries were generated and sequenced. Logistic regression algorithms were employed using 5hmC feature sets combined with physical characteristics of DNA fragments to optimally partition cancer from non-cancer samples. Results: Cross validation of the training model yielded an overall sensitivity of 65.9%,(95% CI, 57.2%–73.9%), early-stage (stage I-II) sensitivity of 57.1% (95% CI, 44%–69.5%) and a specificity of 98%. The model was further validated in a separate, non-overlapping set of blinded and independently processed samples and yielded an early-stage sensitivity of 68.3% (95% CI, 51.9%–81.9%) and a specificity of 96.9% (95% CI, 96.0%–97.6%). Conclusions: Our results demonstrate that plasma-derived cfDNA 5hmC profiles enable the accurate detection of early-stage PaCa, providing a valuable non-invasive tool especially for those individuals at high risk for the disease, including individuals with genetic predisposition and newly diagnosed type 2 diabetes. A larger clinical study (NODMED - NCT05188586) is ongoing and will provide clinical validation for the detection in individuals at high risk for this deadly disease.

A novel CSP C-terminal epitope targeted by an antibody with protective activity against Plasmodium falciparum

Potent and durable vaccine responses will be required for control of malaria caused by Plasmodium falciparum (Pf). RTS,S/AS01 is the first, and to date, the only vaccine that has demonstrated significant reduction of clinical and severe malaria in endemic cohorts in Phase 3 trials. Although the vaccine is protective, efficacy declines over time with kinetics paralleling the decline in antibody responses to the Pf circumsporozoite protein (PfCSP). Although most attention has focused on antibodies to repeat motifs on PfCSP, antibodies to other regions may play a role in protection. Here, we expressed and characterized seven monoclonal antibodies to the C-terminal domain of CSP (ctCSP) from volunteers immunized with RTS,S/AS01. Competition and crystal structure studies indicated that the antibodies target two different sites on opposite faces of ctCSP. One site contains a polymorphic region (denoted α-ctCSP) and has been previously characterized, whereas the second is a previously undescribed site on the conserved β-sheet face of the ctCSP (denoted β-ctCSP). Antibodies to the β-ctCSP site exhibited broad reactivity with a diverse panel of ctCSP peptides whose sequences were derived from field isolates of P. falciparum whereas antibodies to the α-ctCSP site showed very limited cross reactivity. Importantly, an antibody to the β-site demonstrated inhibition activity against malaria infection in a murine model. This study identifies a previously unidentified conserved epitope on CSP that could be targeted by prophylactic antibodies and exploited in structure-based vaccine design.

Pancreatic cancer detection using 5-hydroxymethylation signatures in plasma-derived cell free DNA in high-risk patients with new onset diabetes

539

Background: Pancreatic cancer (PaCa) is the third leading cause of cancer death in the United States despite a low incidence rate. It is often diagnosed when cancer has already metastasized to distant organs. Late diagnosis deprives patients of potentially curative treatments such as surgery and impacts survival rates. People with new onset diabetes (NOD) are at 6-8 fold increased risk for PaCa compared to the general population. Indeed, 0.85% of patients with NOD will be diagnosed with PaCa within 3 years. This population of PaCa patients with NOD constitute 25% of all new pancreatic cancer diagnoses. Surveillance of the NOD population for PaCa presents an opportunity to shift PaCa diagnosis to earlier stage. Methods: Whole blood was obtained from a cohort of 167 PaCa patients and 490 patients with cancers other than PaCa as well as 836 non-cancer controls with and without NOD. Plasma was processed to isolate cfDNA and 5hmC libraries were generated and sequenced. 5hmC data is used to generate models for PaCa detection using Bluestar Genomics’s technology platform. Results: To investigate whether PaCa can be detected in plasma, we interrogated plasma-derived cfDNA hydroxymethylation in PaCa patients and non-cancer controls. Models trained using 5hmC-based biomarkers from cfDNA consistently performed with a mean test sensitivity of 61.1% [95% confidence interval (CI): 35.7% to 82.7%] and a test specificity of 97.6% (CI: 93% to 99.5%) measured across 50 cross validation iterations within the training data set, which was composed of 48.3% early stage (Stages I & II) disease. The final model was trained using all of the training data, yielding 58.4% (CI: 47.5% to 68.8%) sensitivity at 98% (CI: 96.5% to 99.0%) specificity. This model was then tested on an independent test set with 22 PaCa patients (51.7% early stage, 15 of which was NOD) and 123 non-cancer control patients (53 of which were NOD) and yielded a classification performance of 59.1% (CI: 36.4% to 79.3%) sensitivity at 95.9% (CI: 90.8% to 98.7%) specificity. The model performance in the subset of patient cohort with NOD was 53.3% (CI: 26.6% to 78.7%) sensitivity at 94.3% (CI: 84.3% to 98.8%) specificity. Lastly, sensitivity observed on an independent validation set, composed of 56 PaCa and 117 ITTP samples, was 46.4% (CI: 33.0% to 60.2%) with 100% (CI: 96.8 to 100%) specificity. Conclusions: Our results demonstrate PaCa detection in plasma-derived cfDNA using 5hmC profiles. Overall, the model performed consistently between the training and independent validation datasets. A larger clinical study is under development to clinically validate the model described in this study with the goal of identifying occult PaCa within the NOD population in order to enable earlier detection and thus improve patient outcomes.

Pancreatic cancer detection using EpiDetect signatures in plasma-derived cell free DNA in high-risk patients with new onset diabetes

e16265

Background: Pancreatic cancer (PaCa) is the third leading cause of cancer death in the United States despite its low incidence rate, owing to a 5-year survival rate of 10%. It is often asymptomatic in early stage, resulting in the majority of diagnoses occurring when cancer has already metastasized to distant organs. Late diagnosis deprives patients of potentially curative treatments such as surgery and impacts survival rates. Diabetes can be an early symptom of PaCa. Indeed, 25% of PaCa patients had a preceding diabetes diagnosis. Among all people with new onset diabetes (NOD), 0.85% will be diagnosed with PaCa within 3 years, which represents 6-8 fold increased risk for PaCa compared to the general population. Surveillance of the NOD population for PaCa presents an opportunity to shift PaCa diagnosis to earlier stage by finding it sooner. Methods: Whole blood was obtained from a cohort of 117 PaCa patients as well as 800 non-cancer controls with and without NOD. Plasma was processed to isolate cfDNA and 5hmC and low pass whole genome libraries were generated and sequenced. The EpiDetect assay combines 5hmC and whole genome sequencing data and were generated using Bluestar Genomics’s technology platform. Results: To investigate whether PaCa can be detected in plasma, we interrogated plasma-derived cfDNA epigenomic and genomic signal from PaCa patients and non-cancer controls. We first trained stacked ensemble models on PaCa and non-cancer samples utilizing 5hmC, fragmentation and CNV-based biomarkers from cfDNA. These models performed stably with a median of 72.8% sensitivity and 90.1% specificity measured across 25 outer fold iterations using the training data set, which was composed of 50% early stage (Stages I & II) disease. The final binomial ensemble model was trained using all of the training data, yielding an area under the receiver operating characteristic curve (auROC) of 0.9, with 75% sensitivity and 89% specificity. This model was then tested on an independent validation data set from 33 PaCa patients (24 with diabetes, 15 of which was NOD) and 202 non-cancer control patients (76 with diabetes, 51 of which was NOD) and yielded a classification performance auROC of 0.9 with 67% sensitivity at 92% specificity. Lastly, model performance in the subset of patient cohort with NOD only had an auROC of 0.87 with 60% sensitivity at 88% specificity. Conclusions: Our results indicate that 5hmC profiles along with CNV and fragmentation patterns from cfDNA can be used to detect PaCa in plasma-derived cfDNA. Overall, model performance was stable and consistent between the training and independent validation datasets. A larger clinical study is under development to investigate the utility of the model described in this pilot study in identifying occult PaCa within the NOD population, with the aim of shifting diagnosis to early stage and potentially improving patient outcomes.

Genome-wide 5hmC profiles to enable cancer detection and tissue of origin classification in breast, colorectal, lung, ovarian, and pancreatic cancers

3044

Background: Epigenomics assays have recently become popular tools for identification of molecular biomarkers, both in tissue and in plasma. In particular 5-hydroxymethyl-cytosine (5hmC) method, has been shown to enable the epigenomic regulation of gene expression and subsequent gene activity, with different patterns, across several tumor and normal tissues types. In this study we show that 5hmC profiles enable discrete classification of tumor and normal tissue for breast, colorectal, lung ovary and pancreas. Such classification was also recapitulated in cfDNA from patient with breast, colorectal, lung, ovarian and pancreatic cancers. Methods: DNA was isolated from 176 fresh frozen tissues from breast, colorectal, lung, ovary and pancreas (44 per tumor per tissue type and up to 11 tumor tissues for each stage (I-IV)) and up to 10 normal tissues per tissue type. cfDNA was isolated from plasma from 783 non-cancer individuals and 569 cancer patients. Plasma-isolated cfDNA and tumor genomic DNA, were enriched for the 5hmC fraction using chemical labelling, sequenced, and aligned to a reference genome to construct features sets of 5hmC patterns. Results: 5hmC multinomial logistic regression analysis was employed across tumor and normal tissues and identified a set of specific and discrete tumor and normal tissue gene-based features. This indicates that we can classify samples regardless of source, with a high degree of accuracy, based on tissue of origin and also distinguish between normal and tumor status.Next, we employed a stacked ensemble machine learning algorithm combining multiple logistic regression models across diverse feature sets to the cfDNA dataset composed of 783 non cancers and 569 cancers comprising 67 breast, 118 colorectal, 210 Lung, 71 ovarian and 100 pancreatic cancers. We identified a genomic signature that enable the classification of non-cancer versus cancers with an outer fold cross validation sensitivity of 49% (CI 45%-53%) at 99% specificity. Further, individual cancer outer fold cross validation sensitivity at 99% specificity, was measured as follows: breast 30% (CI 119% -42%); colorectal 41% (CI 32%-50%); lung 49% (CI 42%-56%); ovarian 72% (CI 60-82%); pancreatic 56% (CI 46%-66%). Conclusions: This study demonstrates that 5hmC profiles can distinguish cancer and normal tissues based on their origin. Further, 5hmC changes in cfDNA enables detection of the several cancer types: breast, colorectal, lung, ovarian and pancreatic cancers. Our technology provides a non-invasive tool for cancer detection with low risk sample collection enabling improved compliance than current screening methods. Among other utilities, we believe our technology could be applied to asymptomatic high-risk individuals thus enabling enrichment for those subjects that most need a diagnostic imaging follow up.

Structural and biophysical correlation of anti-NANP antibodies with in vivo protection against P. falciparum

The most advanced P. falciparum circumsporozoite protein-based malaria vaccine, RTS,S/AS01 (RTS,S), confers partial protection but with antibody titers that wane relatively rapidly, highlighting the need to elicit more potent and durable antibody responses. Here, we elucidate crystal structures, binding affinities and kinetics, and in vivo protection of eight anti-NANP antibodies derived from an RTS,S phase 2a trial and encoded by three different heavy-chain germline genes. The structures reinforce the importance of homotypic Fab-Fab interactions in protective antibodies and the overwhelmingly dominant preference for a germline-encoded aromatic residue for recognition of the NANP motif. In this study, antibody apparent affinity correlates best with protection in an in vivo mouse model, with the more potent antibodies also recognizing epitopes with repeating secondary structural motifs of type I β- and Asn pseudo 310 turns; such insights can be incorporated into design of more effective immunogens and antibodies for passive immunization.

Diverse Antibody Responses to Conserved Structural Motifs in Plasmodium falciparum Circumsporozoite Protein

Malaria vaccine candidate RTS,S/AS01 is based on the central and C-terminal regions of the circumsporozoite protein (CSP) of P. falciparum. mAb397 was isolated from a volunteer in an RTS,S/AS01 clinical trial, and it protects mice from infection by malaria sporozoites. However, mAb397 originates from the less commonly used VH3-15 germline gene compared to the VH3-30/33 antibodies generally elicited by RTS,S to the central NANP repeat region of CSP. The crystal structure of mAb397 with an NPNA₄ peptide shows that the central NPNA forms a type I β-turn and is the main recognition motif. In most anti-NANP antibodies studied to date, a germline-encoded Trp is used to engage the Pro in NPNA β-turns, but here the Trp interacts with the first Asn. This “conserved” Trp, however, can arise from different germline genes and be located in the heavy or the light chain. Variation in the terminal ψ angles of the NPNA β-turns results in different dispositions of the subsequent NPNA and, hence, different stoichiometries and modes of antibody binding to rsCSP. Diverse protective antibodies against NANP repeats are therefore not limited to a single germline gene response or mode of binding.

•

mAb397 from an RTS,S trial binds NANP repeats of malaria circumsporozoite protein.

•

Protective mAb397 is encoded by the VH3-15 gene that is rare for NANP antibodies.

•

Anti-NANP mAbs bind type I β-turns in three modes using germline-encoded Trp residues.

•

The Trp of mAb397 interacts with Asn instead of Pro in the NPNA type I β-turn.

•

Diversity in NANP immune responses can aid in next-generation vaccine design.

Remote Activation of a Latent Epitope in an Autoantigen Decoded With Simulated B-Factors

Mutants of a catalytically inactive variant of Proteinase 3 (PR3)—iPR3-Val¹⁰³ possessing a Ser195Ala mutation relative to wild-type PR3-Val¹⁰³—offer insights into how autoantigen PR3 interacts with antineutrophil cytoplasmic antibodies (ANCAs) in granulomatosis with polyangiitis (GPA) and whether such interactions can be interrupted. Here we report that iHm5-Val¹⁰³, a triple mutant of iPR3-Val¹⁰³, bound a monoclonal antibody (moANCA518) from a GPA patient on an epitope remote from the mutation sites, whereas the corresponding epitope of iPR3-Val¹⁰³ was latent to moANCA518. Simulated B-factor analysis revealed that the binding of moANCA518 to iHm5-Val¹⁰³ was due to increased main-chain flexibility of the latent epitope caused by remote mutations, suggesting rigidification of epitopes with therapeutics to alter pathogenic PR3·ANCA interactions as new GPA treatments.

Potent antibody lineage against malaria transmission elicited by human vaccination with Pfs25

Transmission-blocking vaccines have the potential to be key contributors to malaria elimination. Such vaccines elicit antibodies that inhibit parasites during their development in Anopheles mosquitoes, thus breaking the cycle of transmission. To date, characterization of humoral responses to Plasmodium falciparum transmission-blocking vaccine candidate Pfs25 has largely been conducted in pre-clinical models. Here, we present molecular analyses of human antibody responses generated in a clinical trial evaluating Pfs25 vaccination. From a collection of monoclonal antibodies with transmission-blocking activity, we identify the most potent transmission-blocking antibody yet described against Pfs25; 2544. The interactions of 2544 and three other antibodies with Pfs25 are analyzed by crystallography to understand structural requirements for elicitation of human transmission-blocking responses. Our analyses provide insights into Pfs25 immunogenicity and epitope potency, and detail an affinity maturation pathway for a potent transmission-blocking antibody in humans. Our findings can be employed to guide the design of improved malaria transmission-blocking vaccines.

Pfs25 is a transmission-blocking vaccine candidate for Plasmodium. Here, McLeod et al. analyze the antibody response to Pfs25 in sera from a clinical trial evaluating a Pfs25 vaccine candidate, identify a potent transmission-blocking antibody and determine recognized epitopes on Pfs25.

Abstract 3966: Mining the cancer immuno-responsome: The identification of functional antitumor antibodies from patients receiving checkpoint inhibitors

Background: The role of B cells and antibodies in anticancer immune responses may correlate with improved prognosis in several types of cancer. Indeed, tumor-reactive antibodies are detected in the blood of cancer patients, tumor-infiltrating B cells have been shown to produce tumor-reactive antibodies, and tumor-reactive antibodies can cause tumor regression in several mouse models. Taken together, these observations support further identification, isolation and characterization of antitumor antibodies from patients demonstrating effective anticancer responses and defining the cognate targets and mechanisms whereby they contribute to tumor control.

Methods: We identified cohorts of patients with nonprogressing metastatic cancer who had received checkpoint immunotherapy and isolated their circulating plasmablasts. Antibody heavy and light chain paired sequences were obtained from individual cells using Atreca's Immune Repertoire Capture (IRCTM) technology. The expressed antibodies were then analyzed for their ability to bind to tumor cells as well as tumor tissue and their ability to mediate antitumor activity was explored in syngeneic mouse tumor models.

Results: Elevated plasmablast levels were observed in individuals with nonprogressing metastatic cancer, and analysis of plasmablast antibody sequences revealed clonal families of B cells that persisted over time with hallmarks of affinity maturation and class switching. We also identified antibody sequences with features common to more than one patient, consistent with convergent antibody selection. In particular, one antibody (AB-213) isolated from a NSCLC patient was found to demonstrate binding to unrelated human tumors as well as the mouse EMT6 tumor. When AB-213 was expressed with a mouse IgG2a constant region the chimeric antibody showed efficacy in vivo by reducing tumor volume and increasing survival in Balb/c mice harboring the syngeneic EMT6 model. Antitumor activity of the chimeric antibody was observed to be dose-dependent when administered as monotherapy or in combination with checkpoint inhibitors. We feel, based on these data, AB-213 could become a very important clinical therapeutic.

Citation Format: Gilson Baia, Amy Manning-Bog, Alexander Scholz, Jeff DeFalco, Michael Harbell, Danhui Zhang, Felix Chu, Beatriz Millare, May Sumi, Patricia Zuno, Judevin Lugar Sapugay, Dongkyoon Kim, Yvonne Leung, Shuwei Jiang, Xiaobin Tang, Kevin Williamson, Xiaomu Chen, Sean Carroll, Christine Dowd, Ish Dhawan, Jonathan Benjamin, Gregg Espiritu Santo, Nicole Haaser, Ngan Nguyen, Eldar Giladi, David Minor, Yann Chong Tan, Jeremy B. Sokolove, Lawrence Steinman, Tito Serafini, Guy Cavet, Norman M. Greenberg, Jacob Glanville, Wayne Volkmuth, Daniel E. Emerling, William H. Robinson. Mining the cancer immuno-responsome: The identification of functional antitumor antibodies from patients receiving checkpoint inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3966.

Abstract 615: Increased somatic hypermutation in the immunoglobulin sequences of melanoma patients who have durable response to checkpoint inhibitor therapy

Patient immune response to tumor can drive profound and durable clinical benefit, but the contribution of antibodies to this benefit is poorly understood. To further our understanding of the role of the humoral response, we compared antibody sequence repertoires of durable responder and non-responder melanoma patients before and during checkpoint inhibitor treatment.

We collected pre- and on-treatment peripheral blood from 26 melanoma patients treated with pembrolizumab (9), ipilimumab (8) or nivolumab+ipilimumab (9). Of the 26 patients, 14 were durable responders (RECIST stable disease, partial response, or complete response for at least 6 months) and 12 were non-responders. We generated natively paired heavy and light chain antibody sequences from individual IgG plasmablasts (CD19+CD20-CD38+CD3-CD14-IgA-IgM-IgD-). Sequences were obtained from a total of 26,725 plasmablasts. These sequences were used to reconstruct lineages (sets of plasmablasts likely derived from a common progenitor B cell) where each lineage shares heavy and light V genes, CDR3 length and 80% sequence similarity in CDR3s.

Durable responders exhibited an increase in somatic hypermutation (SHM) after initiation of treatment. In contrast, no significant change in SHM was observed in non-responders. Higher SHM was also observed in on-treatment repertoires of responders compared with those of non-responders suggesting that T cell checkpoint inhibitors promote activation of humoral immunity in clinical responders. Furthermore, persistent antibody lineages (observed both before and during treatment) showed higher SHM than lineages observed at only a single time point.

Comparison of lineages between patients identified antibodies with high sequence similarity suggesting these antibodies may have arisen from convergent selection, i.e., different patients raising antibodies against shared or similar epitopes. These putative convergent antibodies were enriched in IgG2. IgG2 usage was also higher overall in durable responders than non-responders, even in samples taken prior to treatment. The higher levels of IgG2 observed in responders and in the potentially convergent sets of antibodies suggest that IgG2 antibodies may play a role in effective anti-tumor responses.

In conclusion, our analysis of plasmablast repertoires from melanoma patients suggests that treatment with checkpoint inhibitors promotes SHM in durable responders, and that responder plasmablasts are enriched in IgG2. Moreover, patients make potentially convergent antibodies that are also enriched for the IgG2 subclass. Our observations demonstrate that the humoral immune response is remodeled in patients with anti-tumor responses and support a role for the humoral arm of the immune response in driving clinical benefit in patients treated with checkpoint inhibitors.

Citation Format: Ngan Nguyen, Alusha Mamchak, Mariano Severgnini, Kevin S. Williamson, Xinqi Wu, Elliott F. Drabek, Michael Manos, Xiaomu Chen, Xiaobin Tang, Zoe Amiri, Chantia Carroll, Yvonne Leung, Dongkyoon Kim, Wayne Volkmuth, Norman Greenberg, Daniel Emerling, William H. Robinson, Guy Cavet, F. Stephen Hodi. Increased somatic hypermutation in the immunoglobulin sequences of melanoma patients who have durable response to checkpoint inhibitor therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 615.

Non-progressing cancer patients have persistent B cell responses expressing shared antibody paratopes that target public tumor antigens

There is significant debate regarding whether B cells and their antibodies contribute to effective anti-cancer immune responses. Here we show that patients with metastatic but non-progressing melanoma, lung adenocarcinoma, or renal cell carcinoma exhibited increased levels of blood plasmablasts. We used a cell-barcoding technology to sequence their plasmablast antibody repertoires, revealing clonal families of affinity matured B cells that exhibit progressive class switching and persistence over time. Anti-CTLA4 and other treatments were associated with further increases in somatic hypermutation and clonal family size. Recombinant antibodies from clonal families bound non-autologous tumor tissue and cell lines, and families possessing immunoglobulin paratope sequence motifs shared across patients exhibited increased rates of binding. We identified antibodies that caused regression of, and durable immunity toward, heterologous syngeneic tumors in mice. Our findings demonstrate convergent functional anti-tumor antibody responses targeting public tumor antigens, and provide an approach to identify antibodies with diagnostic or therapeutic utility.

Fractional Third and Fourth Dose of RTS,S/AS01 Malaria Candidate Vaccine: A Phase 2a Controlled Human Malaria Parasite Infection and Immunogenicity Study

BACKGROUND

Three full doses of RTS,S/AS01 malaria vaccine provides partial protection against controlled human malaria parasite infection (CHMI) and natural exposure. Immunization regimens, including a delayed fractional third dose, were assessed for potential increased protection against malaria and immunologic responses.

METHODS

In a phase 2a, controlled, open-label, study of healthy malaria-naive adults, 16 subjects vaccinated with a 0-, 1-, and 2-month full-dose regimen (012M) and 30 subjects who received a 0-, 1-, and 7-month regimen, including a fractional third dose (Fx017M), underwent CHMI 3 weeks after the last dose. Plasmablast heavy and light chain immunoglobulin messenger RNA sequencing and antibody avidity were evaluated. Protection against repeat CHMI was evaluated after 8 months.

RESULTS

A total of 26 of 30 subjects in the Fx017M group (vaccine efficacy [VE], 86.7% [95% confidence interval [CI], 66.8%-94.6%]; P < .0001) and 10 of 16 in the 012M group (VE, 62.5% [95% CI, 29.4%-80.1%]; P = .0009) were protected against infection, and protection differed between schedules (P = .040, by the log rank test). The fractional dose boosting increased antibody somatic hypermutation and avidity and sustained high protection upon rechallenge.

DISCUSSIONS

A delayed third fractional vaccine dose improved immunogenicity and protection against infection. Optimization of the RTS,S/AS01 immunization regimen may lead to improved approaches against malaria.

CLINICAL TRIALS REGISTRATION

NCT01857869.

Antibodies to influenza nucleoprotein cross-react with human hypocretin receptor 2

The sleep disorder narcolepsy is linked to the HLA-DQB1*0602 haplotype and dysregulation of the hypocretin ligand-hypocretin receptor pathway. Narcolepsy was associated with Pandemrix vaccination (an adjuvanted, influenza pandemic vaccine) and also with infection by influenza virus during the 2009 A(H1N1) influenza pandemic. In contrast, very few cases were reported after Focetria vaccination (a differently manufactured adjuvanted influenza pandemic vaccine). We hypothesized that differences between these vaccines (which are derived from inactivated influenza viral proteins) explain the association of narcolepsy with Pandemrix-vaccinated subjects. A mimic peptide was identified from a surface-exposed region of influenza nucleoprotein A that shared protein residues in common with a fragment of the first extracellular domain of hypocretin receptor 2. A significant proportion of sera from HLA-DQB1*0602 haplotype-positive narcoleptic Finnish patients with a history of Pandemrix vaccination (vaccine-associated narcolepsy) contained antibodies to hypocretin receptor 2 compared to sera from nonnarcoleptic individuals with either 2009 A(H1N1) pandemic influenza infection or history of Focetria vaccination. Antibodies from vaccine-associated narcolepsy sera cross-reacted with both influenza nucleoprotein and hypocretin receptor 2, which was demonstrated by competitive binding using 21-mer peptide (containing the identified nucleoprotein mimic) and 55-mer recombinant peptide (first extracellular domain of hypocretin receptor 2) on cell lines expressing human hypocretin receptor 2. Mass spectrometry indicated that relative to Pandemrix, Focetria contained 72.7% less influenza nucleoprotein. In accord, no durable antibody responses to nucleoprotein were detected in sera from Focetria-vaccinated nonnarcoleptic subjects. Thus, differences in vaccine nucleoprotein content and respective immune response may explain the narcolepsy association with Pandemrix.

Technical advances: genome-wide cDNA-AFLP analysis of the Arabidopsis transcriptome

cDNA-AFLP, a technology historically used to identify small numbers of differentially expressed genes, was adapted as a genome-wide transcript profiling method. mRNA levels were assayed in a diverse range of tissues from Arabidopsis thaliana plants grown under a variety of environmental conditions. The resulting cDNA-AFLP fragments were sequenced. By linking cDNA-AFLP fragments to their corresponding mRNAs via these sequences, a database was generated that contained quantitative expression information for up to two-thirds of gene loci in A. thaliana, ecotype Ws. Using this resource, the expression levels of genes, including those with high nucleotide sequence similarity, could be determined in a high-throughput manner merely by comparing cDNA-AFLP profiles with the database. The lengths of cDNA-AFLP fragments inferred from their electrophoretic mobilities correlated well with actual fragment lengths determined by sequencing. In addition, the concentrations of AFLP fragments from single cDNAs were highly correlated, illustrating the validity of cDNA-AFLP as a quantitative, genome-wide, transcript profiling method. cDNA-AFLP profiles were also qualitatively consistent with mRNA profiles obtained from parallel microarray analysis, and with data from previous studies.

Evidence for sequence-independent evolutionary traces in genomics data

Sequence conservation during evolution is the foundation for the functional classification of the enormous number of new protein sequences being discovered in the current era of genome sequencing. Conventional methods to detect homologous proteins are not always able to distinguish between true homologs and false positive hits in the twilight zone of sequence similarity. Several different approaches have been proposed to improve the sensitivity of these methods. Among the most successful are sequence profiles, multi-linked alignment, and threading. However, evolution might offer up other clues about a protein's ancestry that are sequence independent. Here we report the discovery of two such traces of evolution that could potentially be used to help infer the fold of a protein and hence improve the ability to predict the biochemical function. The first such evolutionary trace is a conservation of fold along the genome, i.e. nearby genes tend to share a fold more often than expected by chance alone--a not unexpected observation, but one which holds true even when no pair of genes being examined share appreciable homology. The second such evolutionary trace is, surprisingly, present in expression data: genes that are correlated in expression are more apt to share a fold than two randomly chosen genes. This result is surprising because correlations in expression have previously only been considered useful for determining biological function (e.g. what pathway a particular gene fits into), yet the observed fold enrichment in the expression data permits us to say something about biochemical function since fold corresponds strongly with biochemical function. Again, the fold enrichment observed in the expression data is apparent even when no pair of genes being examined share appreciable homology.

SST: an algorithm for finding near-exact sequence matches in time proportional to the logarithm of the database size

MOTIVATION

Searches for near exact sequence matches are performed frequently in large-scale sequencing projects and in comparative genomics. The time and cost of performing these large-scale sequence-similarity searches is prohibitive using even the fastest of the extant algorithms. Faster algorithms are desired.

RESULTS

We have developed an algorithm, called SST (Sequence Search Tree), that searches a database of DNA sequences for near-exact matches, in time proportional to the logarithm of the database size n. In SST, we partition each sequence into fragments of fixed length called 'windows' using multiple offsets. Each window is mapped into a vector of dimension 4(k) which contains the frequency of occurrence of its component k-tuples, with k a parameter typically in the range 4-6. Then we create a tree-structured index of the windows in vector space, with tree-structured vector quantization (TSVQ). We identify the nearest neighbors of a query sequence by partitioning the query into windows and searching the tree-structured index for nearest-neighbor windows in the database. When the tree is balanced this yields an O(logn) complexity for the search. This complexity was observed in our computations. SST is most effective for applications in which the target sequences show a high degree of similarity to the query sequence, such as assembling shotgun sequences or matching ESTs to genomic sequence. The algorithm is also an effective filtration method. Specifically, it can be used as a preprocessing step for other search methods to reduce the complexity of searching one large database against another. For the problem of identifying overlapping fragments in the assembly of 120 000 fragments from a 1.5 megabase genomic sequence, SST is 15 times faster than BLAST when we consider both building and searching the tree. For searching alone (i.e. after building the tree index), SST 27 times faster than BLAST.

AVAILABILITY

Request from the authors.

Comprehensive gene expression analysis by transcript profiling

After the completion of the genomic sequence of Arabidopsis thaliana, it is now a priority to identify all the genes, their patterns of expression and functions. Transcript profiling is playing a substantial role in annotating and determining gene functions, having advanced from one-gene-at-a-time methods to technologies that provide a holistic view of the genome. In this review, comprehensive transcript profiling methodologies are described, including two that are used extensively by the authors, cDNA-AFLP and cDNA microarraying. Both these technologies illustrate the requirement to integrate molecular biology, automation, LIMS and data analysis. With so much uncharted territory in the Arabidopsis genome, and the desire to tackle complex biological traits, such integrated systems will provide a rich source of data for the correlative, functional annotation of genes.

Pharmaceutical target discovery using Guilt-by-Association: schizophrenia and Parkinson's disease genes

We wish to identify genes associated with disease. To do so, we look for novel genes whose expression patterns mimic those of known disease-associated genes, a method we call Guilt-by-Association (GBA). GBA uses a combinatoric measure of association that provides superior results to those from correlation measures used in previous expression analyses. Using GBA, we have examined the expression of 40,000 human genes in 522 cDNA libraries, and have identified several hundred genes associated with known cancer, inflammation, steroid-synthesis, insulin-synthesis, neurotransmitter processing, matrix remodeling and other disease genes. The majority of the genes thus discovered show no significant sequence similarity to known genes, and thus could not have been identified by homology searches. We present here an example of the discovery of five genes associated with schizophrenia and Parkinson's disease. Of the 40,000 most-abundant human genes, these five genes are the most closely linked to the known disease genes, and thus are prime targets for pharmaceutical intervention.

Prediction of gene function by genome-scale expression analysis: prostate cancer-associated genes

We wish to identify genes associated with disease. To do so, we look for novel genes whose expression patterns mimic those of known disease-associated genes, using a method we call Guilt-by-Association (GBA), on the basis of a combinatoric measure of association. Using GBA, we have examined the expression of 40,000 human genes in 522 cDNA libraries, and have discovered several hundred previously unidentified genes associated with cancer, inflammation, steroid-synthesis, insulin-synthesis, neurotransmitter processing, matrix remodeling, and other disease processes. The majority of the genes thus discovered show no sequence similarity to known genes, and thus could not have been identified by homology searches. We present here an example of the discovery of eight genes associated with prostate cancer. Of the 40,000 most-abundant human genes, these 8 are the most closely linked to the known diagnostic genes, and thus are prime targets for pharmaceutical research.