Adaptive adversarial neural networks for the analysis of lossy and domain-shifted datasets of medical images.

In machine learning for image-based medical diagnostics, supervised convolutional neural networks are typically trained with large and expertly annotated datasets obtained using high-resolution imaging systems. Moreover, the network's performance can degrade substantially when applied to a dataset with a different distribution. Here, we show that adversarial learning can be used to develop high-performing networks trained on unannotated medical images of varying image quality. Specifically, we used low-quality images acquired using inexpensive portable optical systems to train networks for the evaluation of human embryos, the quantification of human sperm morphology and the diagnosis of malarial infections in the blood, and show that the networks performed well across different data distributions. We also show that adversarial learning can be used with unlabelled data from unseen domain-shifted datasets to adapt pretrained supervised networks to new distributions, even when data from the original distribution are not available. Adaptive adversarial networks may expand the use of validated neural-network models for the evaluation of data collected from multiple imaging systems of varying quality without compromising the knowledge stored in the network.

T cell-tropic HIV efficiently infects alveolar macrophages through contact with infected CD4+ T cells.

Alveolar macrophages (AMs) are critical for defense against airborne pathogens and AM dysfunction is thought to contribute to the increased burden of pulmonary infections observed in individuals living with HIV-1 (HIV). While HIV nucleic acids have been detected in AMs early in infection, circulating HIV during acute and chronic infection is usually CCR5 T cell-tropic (T-tropic) and enters macrophages inefficiently in vitro. The mechanism by which T-tropic viruses infect AMs remains unknown. We collected AMs by bronchoscopy performed in HIV-infected, antiretroviral therapy (ART)-naive and uninfected subjects. We found that viral constructs made with primary HIV envelope sequences isolated from both AMs and plasma were T-tropic and inefficiently infected macrophages. However, these isolates productively infected macrophages when co-cultured with HIV-infected CD4+ T cells. In addition, we provide evidence that T-tropic HIV is transmitted from infected CD4+ T cells to the AM cytosol. We conclude that AM-derived HIV isolates are T-tropic and can enter macrophages through contact with an infected CD4+ T cell, which results in productive infection of AMs. CD4+ T cell-dependent entry of HIV into AMs helps explain the presence of HIV in AMs despite inefficient cell-free infection, and may contribute to AM dysfunction in people living with HIV.

SARS-CoV-2 viral load is associated with increased disease severity and mortality.

The relationship between SARS-CoV-2 viral load and risk of disease progression remains largely undefined in coronavirus disease 2019 (COVID-19). Here, we quantify SARS-CoV-2 viral load from participants with a diverse range of COVID-19 disease severity, including those requiring hospitalization, outpatients with mild disease, and individuals with resolved infection. We detected SARS-CoV-2 plasma RNA in 27% of hospitalized participants, and 13% of outpatients diagnosed with COVID-19. Amongst the participants hospitalized with COVID-19, we report that a higher prevalence of detectable SARS-CoV-2 plasma viral load is associated with worse respiratory disease severity, lower absolute lymphocyte counts, and increased markers of inflammation, including C-reactive protein and IL-6. SARS-CoV-2 viral loads, especially plasma viremia, are associated with increased risk of mortality. Our data show that SARS-CoV-2 viral loads may aid in the risk stratification of patients with COVID-19, and therefore its role in disease pathogenesis should be further explored.

Publisher Correction: Evidence that CD32a does not mark the HIV-1 latent reservoir.

In this Brief Communications Arising Comment, the first three authors (Osuna, Lim and Kublin) should have been listed as equally contributing authors; this has been corrected online.

Rapid, label-free CD4 testing using a smartphone compatible device.

The most recent guidelines have called for a significant shift towards viral load testing for HIV/AIDS management in developing countries; however point-of-care (POC) CD4 testing still remains an important component of disease staging in multiple developing countries. Advancements in micro/nanotechnologies and consumer electronics have paved the way for mobile healthcare technologies and the development of POC smartphone-based diagnostic assays for disease detection and treatment monitoring. Here, we report a simple, rapid (30 minutes) smartphone-based microfluidic chip for automated CD4 testing using a small volume (30 µL) of whole blood. The smartphone-based device includes an inexpensive (<$5) cell phone accessory and a functionalized disposable microfluidic device. We evaluated the performance of the device using spiked PBS samples and HIV-infected and uninfected whole blood, and compared the microfluidic chip results with the manual analysis and flow cytometry results. Through t-tests, Bland-Altman analyses, and regression tests, we have shown a good agreement between the smartphone-based test and the manual and FACS analysis for CD4 count. The presented technology could have a significant impact on HIV management in developing countries through providing a reliable and inexpensive POC CD4 testing.

HIV Impairs Lung Epithelial Integrity and Enters the Epithelium to Promote Chronic Lung Inflammation.

Several clinical studies show that individuals with HIV are at an increased risk for worsened lung function and for the development of COPD, although the mechanism underlying this increased susceptibility is poorly understood. The airway epithelium, situated at the interface between the external environment and the lung parenchyma, acts as a physical and immunological barrier that secretes mucins and cytokines in response to noxious stimuli which can contribute to the pathobiology of chronic obstructive pulmonary disease (COPD). We sought to determine the effects of HIV on the lung epithelium. We grew primary normal human bronchial epithelial (NHBE) cells and primary lung epithelial cells isolated from bronchial brushings of patients to confluence and allowed them to differentiate at an air- liquid interface (ALI) to assess the effects of HIV on the lung epithelium. We assessed changes in monolayer permeability as well as the expression of E-cadherin and inflammatory modulators to determine the effect of HIV on the lung epithelium. We measured E-cadherin protein abundance in patients with HIV compared to normal controls. Cell associated HIV RNA and DNA were quantified and the p24 viral antigen was measured in culture supernatant. Surprisingly, X4, not R5, tropic virus decreased expression of E-cadherin and increased monolayer permeability. While there was some transcriptional regulation of E-cadherin, there was significant increase in lysosome-mediated protein degradation in cells exposed to X4 tropic HIV. Interaction with CXCR4 and viral fusion with the epithelial cell were required to induce the epithelial changes. X4 tropic virus was able to enter the airway epithelial cells but not replicate in these cells, while R5 tropic viruses did not enter the epithelial cells. Significantly, X4 tropic HIV induced the expression of intercellular adhesion molecule-1 (ICAM-1) and activated extracellular signal-regulated kinase (ERK). We demonstrate that HIV can enter airway epithelial cells and alter their function by impairing cell-cell adhesion and increasing the expression of inflammatory mediators. These observed changes may contribute local inflammation, which can lead to lung function decline and increased susceptibility to COPD in HIV patients.

CCR5 Expression Levels in HIV-Uninfected Women Receiving Hormonal Contraception.

Human immunodeficiency virus (HIV) infectivity increases as receptor/coreceptor expression levels increase. We determined peripheral CD4, CCR5, and CXCR4 expression levels in HIV-uninfected women who used depot medroxyprogesterone acetate (DMPA; n = 32), the levonorgestrel-releasing intrauterine device (LNG-IUD; n = 27), oral contraceptive pills (n = 32), or no hormonal contraception (n = 33). The use of LNG-IUD increased the proportion of CD4(+) and CD8(+) T cells that expressed CCR5; increases in the magnitude of T-cell subset CCR5 expression were observed with DMPA and LNG-IUD use (P < .01 for all comparisons). LNG-IUD and, to a lesser extent, DMPA use were associated with increased peripheral T-cell CCR5 expression.

Acute on-chip HIV detection through label-free electrical sensing of viral nano-lysate.

Development of portable biosensors has broad applications in environmental monitoring, clinical diagnosis, public health, and homeland security. There is an unmet need for pathogen detection at the point-of-care (POC) using a fast, sensitive, inexpensive, and easy-to-use method that does not require complex infrastructure and well-trained technicians. For instance, detection of Human Immunodeficiency Virus (HIV-1) at acute infection stage has been challenging, since current antibody-based POC technologies are not effective due to low concentration of antibodies. In this study, we demonstrated for the first time a label-free electrical sensing method that can detect lysed viruses, i.e. viral nano-lysate, through impedance analysis, offering an alternative technology to the antibody-based methods such as dipsticks and Enzyme-linked Immunosorbent Assay (ELISA). The presented method is a broadly applicable platform technology that can potentially be adapted to detect multiple pathogens utilizing impedance spectroscopy for other infectious diseases including herpes, influenza, hepatitis, pox, malaria, and tuberculosis. The presented method offers a rapid and portable tool that can be used as a detection technology at the POC in resource-constrained settings, as well as hospital and primary care settings.

Vicriviroc resistance decay and relative replicative fitness in HIV-1 clinical isolates under sequential drug selection pressures.

We previously described an HIV-1-infected individual who developed resistance to vicriviroc (VCV), an investigational CCR5 antagonist, during 28 weeks of therapy (Tsibris AM et al., J. Virol. 82:8210-8214, 2008). To investigate the decay of VCV resistance mutations, a standard clonal analysis of full-length env (gp160) was performed on plasma HIV-1 samples obtained at week 28 (the time of VCV discontinuation) and at three subsequent time points (weeks 30, 42, and 48). During 132 days, VCV-resistant HIV-1 was replaced by VCV-sensitive viruses whose V3 loop sequences differed from the dominant pretreatment forms. A deep-sequencing analysis showed that the week 48 VCV-sensitive V3 loop form emerged from a preexisting viral variant. Enfuvirtide was added to the antiretroviral regimen at week 30; by week 48, enfuvirtide treatment selected for either the G36D or N43D HR-1 mutation. Growth competition experiments demonstrated that viruses incorporating the dominant week 28 VCV-resistant env were less fit than week 0 viruses in the absence of VCV but more fit than week 48 viruses. This week 48 fitness deficit persisted when G36D was corrected by either site-directed mutagenesis or week 48 gp41 domain swapping. The correction of N43D, in contrast, restored fitness relative to that of week 28, but not week 0, viruses. Virus entry kinetics correlated with observed fitness differences; the slower entry of enfuvirtide-resistant viruses corrected to wild-type rates in the presence of enfuvirtide. These findings suggest that while VCV and enfuvirtide select for resistance mutations in only one env subunit, gp120 and gp41 coevolve to maximize viral fitness under sequential drug selection pressures.

Differential use of CCR5 by HIV-1 clinical isolates resistant to small-molecule CCR5 antagonists.

How HIV-1 resistant to small-molecule CCR5 antagonists uses the coreceptor for entry has been studied in a limited number of isolates. We characterized dependence on the N terminus (NT) and the second extracellular loop (ECL2) of CCR5 of three vicriviroc (VCV)-resistant clinical isolates broadly cross-resistant to other CCR5 antagonists. Pseudoviruses were constructed to assess CCR5 use by VCV-sensitive and -resistant envelopes of subtype B and C viruses. We determined the extent of entry inhibition by monoclonal antibodies (MAbs) directed against the NT and ECL2 in the presence and absence of VCV and the capacity of these pseudoviruses to use CCR5 mutants that contained scanning alanine substitutions in the CCR5 NT and ECL2 domains. Sensitive and resistant viruses were completely and competitively inhibited by the ECL2-specific MAb 2D7, whereas the NT-specific MAb CTC5 led to partial noncompetitive inhibition. VCV-resistant clones showed greater sensitivity to 2D7 than VCV-sensitive clones, but in the presence of saturating VCV concentrations, the 2D7 susceptibilities of two VCV-resistant viruses were similar to that of VCV-sensitive virus. The entry of VCV-sensitive and -resistant isolates was impaired to differing degrees by alanine mutations in CCR5; substitutions in NT had the greatest effect on viral entry. HIV-1 clinical isolates broadly resistant to CCR5 antagonists demonstrated significant heterogeneity in their use of CCR5. This heterogeneity makes it difficult to draw general conclusions about the relationship between patterns of CCR5 antagonist resistance and the use of specific CCR5 domains for entry.

HIV-1 clinical isolates resistant to CCR5 antagonists exhibit delayed entry kinetics that are corrected in the presence of drug.

HIV CCR5 antagonists select for env gene mutations that enable virus entry via drug-bound coreceptor. To investigate the mechanisms responsible for viral adaptation to drug-bound coreceptor-mediated entry, we studied viral isolates from three participants who developed CCR5 antagonist resistance during treatment with vicriviroc (VCV), an investigational small-molecule CCR5 antagonist. VCV-sensitive and -resistant viruses were isolated from one HIV subtype C- and two subtype B-infected participants; VCV-resistant isolates had mutations in the V3 loop of gp120 and were cross-resistant to TAK-779, an investigational antagonist, and maraviroc (MVC). All three resistant isolates contained a 306P mutation but had variable mutations elsewhere in the V3 stem. We used a virus-cell ß-lactamase (BlaM) fusion assay to determine the entry kinetics of recombinant viruses that incorporated full-length VCV-sensitive and -resistant envelopes. VCV-resistant isolates exhibited delayed entry rates in the absence of drug, relative to pretherapy VCV-sensitive isolates. The addition of drug corrected these delays. These findings were generalizable across target cell types with a range of CD4 and CCR5 surface densities and were observed when either population-derived or clonal envelopes were used to construct recombinant viruses. V3 loop mutations alone were sufficient to restore virus entry in the presence of drug, and the accumulation of V3 mutations during VCV therapy led to progressively higher rates of viral entry. We propose that the restoration of pre-CCR5 antagonist therapy HIV entry kinetics drives the selection of V3 loop mutations and may represent a common mechanism that underlies the emergence of CCR5 antagonist resistance.

SHIV-162P3 infection of rhesus macaques given maraviroc gel vaginally does not involve resistant viruses.

Maraviroc (MVC) gels are effective at protecting rhesus macaques from vaginal SHIV transmission, but breakthrough infections can occur. To determine the effects of a vaginal MVC gel on infecting SHIV populations in a macaque model, we analyzed plasma samples from three rhesus macaques that received a MVC vaginal gel (day 0) but became infected after high-dose SHIV-162P3 vaginal challenge. Two infected macaques that received a placebo gel served as controls. The infecting SHIV-162P3 stock had an overall mean genetic distance of 0.294±0.027%; limited entropy changes were noted across the envelope (gp160). No envelope mutations were observed consistently in viruses isolated from infected macaques at days 14-21, the time of first detectable viremia, nor selected at later time points, days 42-70. No statistically significant differences in MVC susceptibilities were observed between the SHIV inoculum (50% inhibitory concentration [IC(50)] 1.87 nM) and virus isolated from the three MVC-treated macaques (MVC IC(50) 1.18 nM, 1.69 nM, and 1.53 nM, respectively). Highlighter plot analyses suggested that infection was established in each MVC-treated animal by one founder virus genotype. The expected Poisson distribution of pairwise Hamming Distance frequency counts was observed and a phylogenetic analysis did not identify infections with distinct lineages from the challenge stock. These data suggest that breakthrough infections most likely result from incomplete viral inhibition and not the selection of MVC-resistant variants.

Xenotropic murine leukemia virus-related virus prevalence in patients with chronic fatigue syndrome or chronic immunomodulatory conditions.

We investigated the prevalence of xenotropic murine leukemia virus-related virus (XMRV) among 293 participants seen at academic hospitals in Boston, Massachusetts. Participants were recruited from the following 5 groups of patients: chronic fatigue syndrome (n = 32), human immunodeficiency virus infection (n = 43), rheumatoid arthritis (n = 97), hematopoietic stem-cell or solid organ transplant (n = 26), or a general cohort of patients presenting for medical care (n = 95). XMRV DNA was not detected in any participant samples. We found no association between XMRV and patients with chronic fatigue syndrome or chronic immunomodulatory conditions.

Evolution of CCR5 antagonist resistance in an HIV-1 subtype C clinical isolate.

OBJECTIVES: We previously reported vicriviroc (VCV) resistance in an HIV-infected subject and used deep sequencing and clonal analyses to track the evolution of V3 sequence forms over 28 weeks of therapy. Here, we test the contribution of gp120 mutations to CCR5 antagonist resistance and investigate why certain minority V3 variants emerged as the dominant species under drug pressure. METHODS: Nineteen site-directed HIV-1 mutants were generated that contained gp120 VCV resistance mutations. Viral sensitivities to VCV, maraviroc, TAK-779, and HGS004 were determined. RESULTS: Three patterns of susceptibilities were observed as follows: sigmoid inhibition curves with 50% inhibitory concentration similar to pretreatment virus [07J-week 0 (W0)], single mutants with decreased 50% inhibitory concentrations compared with 07J-W0, and mutants that contained ≥5 of 7 VCV resistance mutations with flattened inhibition curves and decreased or negative percent maximal inhibition. Substitutions such as S306P, which sensitized virus to CCR5 antagonists when present as single mutations, were not detected in the baseline virus population but were necessary for maximal resistance when incorporated into V3 backbones that included preexisting VCV resistance mutations. CONCLUSIONS: CCR5 antagonist resistance was reproduced only when a majority of V3 mutations were present. Minority V3 loop variants may serve as a scaffold upon which additional mutations lead to complete VCV resistance.