Cross-reactivity to gut microbiota could explain failure of HIV vaccine

More than 30 years since its discovery and there is still no FDA licensed HIV vaccine. It is not even entirely clear if the failure to develop effective HIV vaccine has really anything to do with frequent HIV antigenic shift. Flu virus, for example, also undergo quite frequent antigenic shift and still there is Flu vaccines with 60-90% efficacy.   

New study published this week in journal Science suggested new explanation for frequent failure with HIV vaccine(s). It appears that cross-reactivity between gp41 component of HIV vaccine and endogenous microbiota antigen (component of E. coli RNA polymerase) prevented development of neutralizing gp120-specific antibody titre.

The authors have analysed Ab repertoire to HIV-1 DNA prime, recombinant Adenovirus Type 5 (rAd5) boost vaccine. They found that 93% of Env specific antibodies derived from sorted memory B cells were directed against non-neutralizing gp41 antigen. 

Interestingly, majority of those gp41-specific antibodies utilized polyreactive, innate-like IGHV1-69 variable segment. VH1-69 locus is involved in Ab repertoire directed to Flu stem region. But unlike Flu specific VH1-69 Abs, gp41-specific VH1-69 Abs were made of allele variants with Leucine substitution at position 54 in HCDR2.

Finally, analysis of antigen specificity of individual gp41 mAbs revealed high level of polyreactivity towards commensal and self antigens.

In summary, these results points to the most important aspect of immune system, namely that strength and specificity of immune response is controlled by environmental antigens, including microflora antigens. It is my opinion that future vaccine testing would require incorporation of cross-reactivity tests against wide range of environmental antigens to select the most effective immunogens.  

David Usharauli

    

   

Exposure to premature inflammatory milieu inhibits antigen receptor signaling in CD4 T cells

It is believed that proper T cell activation require sequential signaling through Signal 1 (e.g. CD3), Signal 2 (e.g. CD28), Signal 3 (e.g. IL-12 or type I IFN induced inflammation). However, occasionally, for example during autoimmune inflammation or immunotherapy, T cells could be exposed to "out-of-order" Signal 3 without Signal 1. So what are the consequences for such exposure for T cells?

New paper in Immunity has provided some answers. Using human T cells derived from patients treated with high dose of IL-2, the authors found that CD4 T cells, but not CD8 T cells, displayed diminished antigen-specific proliferation in mixed MLR or ConA stimulation assays.

Similar results were obtained with human CD4 T cells pre-incubated in vitro with IL-2 prior to MLR assay (that tests Signal 1 activity). One caveat is that the authors have used linear scale to "highlight" the difference in proliferation of treated T cells. However, the proliferation is a geometrical function and ordinarily should be depicted in logarithmic scale.

Next, the authors found that inhibition of Signal 1 by inflammatory milieu was short-lived and quickly recovered if inflammatory milieu was removed.

Similar quick recovery was seen in vivo with mouse OVA-specific OT-II CD4 T cells treated anti-CD40 antibody and IL-2 to mimic inflammatory milieu and later harvested at different time points post treatment.

Mechanistically, the authors showed that Socs3 was selectively up-regulated in CD4 T cells upon exposure to Signal 3 and inhibited STAT5 signaling in CD4 T cells.

In summary, this very simple paper raises very interesting points regarding how T cells computes external signals. As the authors noted Signal 3 could exist in conditions such as systemic autoimmune inflammation and sepsis, or during immunotherapy. How would these conditions affect activation of CD4 T cells?

David Usharauli


  

Memory T cells can eliminate virus from the brain without tissue damage

Dealing with viral infection in the brain (CNS) is a big challenge to immune system. Unlike other tissues, even minor local inflammation in CNS could make the host unfit to survive. So what immune system can do?

New paper in Journal of Experimental Medicine (JEM) suggested that T cells are using non-cytopathic, non-inflammatory effector signalling to clear the brain from the persisting virus.

Using neonatal brain infection model (carrier mice), the authors showed that adoptive transfer of donor memory T cells into carrier mice did not induce brain tissue damage.

The authors found that memory T cell immunotherapy of carrier mice was associated with limited inflammatory cytokine release and minimal brain tissue damage.

Examination of brain tissue in carrier mice showed that adoptive anti-viral T cell therapy induced brain-wide modification of virus-infected microglia population (CD11c up-regulation).  

Further experiments showed that memory T cells activated IFN-γ mediated STAT1 signaling in the brain.
 

Finally, the authors found that memory T cells cleared virus from infected microglia using non-cytopathic IFN-γ / STAT1 pathway.

In summary, these results suggest that immune system can employ tissue-specific defense mechanisms, as proposed by Matzinger and Kamala.

David Usharauli

6 antigen-specific engineered T cells per µL of blood is needed to protect against multiple myeloma

CAR-T cells express B cell receptors on T cell body. They can target cells expressing proteins. This imposes some limitations on CAR-T cells. Tumors may not express mutated surface proteins readily distinguishable from normal variants. Such differences, however, may be visible at the peptide level. This therapeutic gap could be filled with TCR-engineered T cells targeting tumor-associated MHC+peptide combinations with high affinity and avidity.

New paper in Nature Medicine has provided some early results based on such protocols with NY-ESO-1-specific TCR engineered T cells in multiple myeloma patients. The senior author on this study is Carl June from Upenn whose work on engineered T cells reinvigorated tumor immunotherapy field in recent years.   

Initially, the authors showed that myeloma antigen-specific autologous TCR engineered T cells could persist up to 1 year or more after infusion in patients (though quantification method used here could be unreliable).

Next, the authors showed that lentiviral vector-transduced T cells could migrate and then eliminate myeloma cells from the bone marrow. However, the authors observed that in one patient TCR engineered T cells failed to control myeloma cells once it escaped bone marrow to other tissues. This is concerning. 

In addition, TCR affinity or its expression level in TCR engineered T cells goes dramatically down within 1 year of infusion. This is concerning too.

Finally, the authors claimed that no patients relapsed who had more than 6 antigen-specific T cells per µL of blood. However, close examination of the data also showed that some patients were progression free even though they had less 6 antigen-specific T cells per µL of blood. This implies that T cell number per se  is not a decisive factor here.

In summary, this study showed that close to 70% of patients had a complete or near complete response after TCR-engineered T cell infusion, with minimal associated safety issues.

David Usharauli           

Antigen-specific NK cell memory

Do NK cells have memory? New study in Nature Immunology suggests they do. 

This study was done on rhesus macaques. Purified NK cells were sorted from spleen, liver or blood. Autologous, virus infected or antigen-pulsed DCs were used as targets. Control DCs were either uninfected or mismatched antigen-pulsed.

Initially, the authors observed that purified spleen (c) or hepatic (d) NK cells from SHIV-infected hosts showed cytotoxic response to Gag and Env, but not OVA pulsed autologous DCs.

Similarly, purified splenic NK cells derived from SIV-infected hosts showed antigen-specific cytotoxicity (a, b). Interestingly, blood derived NK cells showed no cytotoxic activity against antigen-pulsed DCs (c).   

Finally, the authors showed that splenic (c) and hepatic (d) NK cells derived from hosts vaccinated 5 years earlier displayed vaccine-matching antigen-specific DCs lysis. Here too, blood derived NK cells showed minimal antigen-pulsed DCs lysis (e).

Additional results indicated that NK-mediated lysis of antigen-pulsed DCs were mediated via NKG2C or NKG2A pathways. These NK receptors recognize MHC class I molecules loaded with leader peptides derived from other endogenous MHC molecules. However, it is not clear how this mechanism confers exogenous antigen-specific memory to NK cells.

In summary, these results suggest that NK cells could mount antigen-specific memory response. 

David Usharauli 

TH17-derived IL-26 punches holes in bacteria

T_H17 cells have been implicated in host's protection against extracellular bacteria. It is thought that T_H17 cells recruit neutrophils and other granulocytes to the site of bacterial infection who do the job.

However, a new study in Nature Immunology provided evidence that human T_H17 cells can directly eliminate extracellular bacteria via release of IL-26, a cationic peptide that punches holes in bacterial membrane.  

This is actually very well done study. Initially the authors showed that recombinant human IL-26 (rhIL-26) had a direct bactericidal activity in vitro.

Next, the authors showed that rhIL-26 had bactericidal activity in vivo as well.

The authors found that in humans IL-26 expression was restricted to T cells and that out of T cells, T_H17 cells expressed high level of IL-26 and that T_H17 cell supernatant showed IL-26 dependent bactericidal activity.

In addition to its direct bactericidal activity, rhIL-26 could activate human plasmacytoid DCs (pDCs) when combined with dying bacterial DNA.

Interestingly, the authors showed that rhIL-26 could also combine with human DNA and activate pDCs and to a lesser extent monocytes.

Further experiments revealed that rhIL-26 + human DNA conjugate were taken up by pDCs and could signal through TLR9.

Finally, the authors showed that T_H17 cells could activate human pDCs to produce IFN-α via IL-26 + DNA conjugate formation.

In summary, these results provided mechanistic model for direct bactericidal activity of T_H17 cells against extracellular bacteria. In addition, results with IL-26 + hDNA conjugates and pDCs activation revealed how T_H17 cells could be involved in amplification of immune response during autoimmune conditions.    

David Usharauli

Human long-lived antibody memory to antigens resides in CD19-CD138+CD38+ bone marrow subset

Development of protective long-lived antibody titre to antigens is a hallmark of successful vaccination. It is widely accepted that long-living plasma cells (LLPCs) residing in bone marrow are responsible for maintaining serum antibody titre. However, almost nothing is known how LLPCs are generated.

New research published in journal Immunity made observation that human LLPC population were selectively enriched in CD19^-CD138⁺CD38⁺ population in bone marrow.

Initially the authors have sorted bone marrow populations enriched in plasmablast/ plasma cell markers from healthy volunteers. All 4 populations were making antibodies.

However, when analyzed for specific antibody production, the authors found that CD19^-CD138⁺CD38⁺ subset was selectively enriched in antibody specificity against antigens encountered several years ago (e.g. mumps, measles, tetanus vaccine). In contrast, Flu-specific antibodies were secreted by several subsets.

To confirm this observation, the authors purified antigen-specific IgG from older volunteers and tested their origin using complementary mRNA sequences in subsets. This confirmed that long-lived antibody memory resided in CD19^-CD138⁺CD38⁺ subset.

In summary, these results indicates that LLPCs in humans are residing in specialized bone marrow subset.

This is a observation study. It does not explain what signals would enhance CD19^-CD138⁺CD38⁺ subset generation. One weakness of this paper is the fact that the authors did not follow LLPCs generation against naive (novel) antigens. Such experiment would have provided separate validation to their hypothesis.

David Usharauli

Immunotherapy of Alzheimer-like cis-tauopathy after repetitive head injury

Alzheimer's disease is a neurodegenerative disorder of unknown etiology. Currently there is no cure. One reason is that we don't know much about its pathophysiology. Two major hypotheses dominate the field: β-amyloid and tau hypothesis. In both models, aggregation and deposition of suspected proteins (β-amyloid or tau) causes neuronal death leading to loss of cognitive functions.

A new paper published in journal Nature belongs to tau model. In this research article, the authors showed that cis, but not trans form of phospho-tau (P-tau) protein contributes to development of Alzheimer-like tauopathy during traumatic brain injury that could be reversed by anti-cis antibody immunotherapy.

Here, using mouse model of traumatic head injury, the authors showed that chronic repetitive head injury caused widespread cis P-tau appearance in the brain.

Next, the authors showed that brain lysates prepared from chronically injured brain tissue could induce in vitro neuronal death that could be prevented by anti-cis antibody.

Similarly, the authors showed anti-cis antibody could reverse neuronal cell death following hypoxia or serum deprivation.

More importantly, mice treated in vivo with anti-cis antibody showed significant reduction of tau protein expression in brain tissue after brain injury.

Mechanistically, the authors hypothesized that anti-cis antibody are taken up by neurons via Fcγ receptors, internalized and interact with intracellular FcR TRIM21 that leads to tau protein detection and its elimination by proteasomal degradation.

However, it is not clear how anti-cis antibody is able to reach neurons, to begin with. BBB is not easily accessible to antibodies.

David Usharauli



     

Foxp3+ Treg-derived IL-10 stabilizes anti-viral CD8 T cell memory formation

CD8 T cells can eliminate infected or tumor cells. Vaccine formulation that could improve CD8 T cell effector function and memory formation would be highly beneficial against intracellular infections such as HIV, TB, malaria or against tumors.

New study in Nature Immunology provided evidence that Foxp3⁺ T_reg-derived IL-10 modulates CD8 T_cm memory formation.

Initially, the authors showed the proportion of central memory CD8 T cells (T_cm) generated after acute viral infection in IL-10KO mice were significantly reduced. No data are shown about actual T_cm numbers, though. No data are shown about actual viral titre or its clearance, either. Not cool.

Interestingly, presence of IL-10 was only relevant between day 8-15 post-infection (during CD8 T effector cell contraction phase). Here the authors have used PBS, rather than isotype antibody, as a control for anti-IL-10 antibody blockade. Not cool.

Next, the authors found that it was IL-10 derived from Foxp3⁺ T_reg cells that influenced T_cm formation.

Conditional ablation of Foxp3⁺ T_reg cells during early stages of infection had the same effect.

More importantly, viral antigen-specific CD8 memory T cells harvested from IL-10 fl/fl Foxp3-Cre mice (that selectively lacks IL-10 in Foxp3⁺ T_regs) showed reduced survival and reduced protection upon secondary antigen exposure in naive recipients.

Mechanistically, the authors hypothesized that hyper-inflammatory milieu found in IL-10 fl/fl Foxp3-Cre hosts at day 15 post viral infection could contribute to reduced T_cm formation. However, the authors did not try to actually test this hypothesis.

Finally, adoptive transfer of WT Foxp3⁺ T_reg cells at day 8 post infection could rescue T_cm phenotype.

In summary, these results suggest that at some point (day 8-15) post viral infection Foxp3⁺ T_reg-derived IL-10 reduces overall host's activation status and enables T_cm phenotype formation.

David Usharauli

Nucleoprotein in Flu vaccine mimics brain protein and may cause autoimmune sleepiness

Narcolepsy is a chronic neurological condition characterized by excessive daytime sleepiness. It causes are unknown. It is suggested that deficiency or inhibition of endogenous brain peptide called hypocretin (orexin) and produced by hypothalamus could recapitulate narcolepsy's symptoms.

New study in Science Translational Medicine suggested that immune cross-reactivity between Flu vaccine component and brain's hypocretin receptor could have caused autoantibody mediated narcolepsy following 2009 Flu H1N1.

First, the authors showed that nucleoprotein (NP) from flu vaccine Pandemrix (that caused narcolepsy) and NP from flu vaccine Focetria (that did not cause narcolepsy) have sequence similarity to human hypocretin receptor. These two NPs differed in only 1 amino acid from each other (isoleucin vs. methionine).

Next, the authors showed that serum samples from Pandermrix cohort, but not from Focetria cohort, reacted with human hypocretin receptor in a sensitive cell-based assay (note high reactivity in healthy Finish group). 

Importantly, the authors showed that this immune reactivity of Pandemprix vaccinated serum samples to human hypocretin receptor could be inhibited by short Flu NP peptides  (both from Pandemprix and Focetria) and or by hypocretin receptor (positive control). It suggested that at the antibody level, there is no difference between NP from Pandemrix or Focetria. The authors could not analyze T cell part.   

In summary, the authors suggested the following hypothesis: high level of NP protein in Pandemrix flu vaccine (compared to Focetria) could have induced robust anti-NP antibody titre that later "somehow" have managed to cross human brain blood barrier (BBB) and caused autoimmune narcolepsy is HLA susceptible individuals. However, it is not clear how autoantibody could have got access to hypothalamic cells expressing hypocretin receptors.

Vaccine are important for human health. However, human HLA diversity and natural mimicry at the protein and peptide level could produce autoimmunity in certain individuals.  

David Usharauli

Serum level of β2-microglobulin, a component of HLA class I system, may predict age-related decline in cognitive function

Ageing-associated decline, especially cognitive decline, is a social-medical condition that would directly affect overall societal wellbeing. It is still widely unmet biomedical challenge. In general, medical conditions related to brain functions are the least understood topics in medicine.

Recently journal Nature Medicine has published an article suggesting that high level of β2-microglobulin, a component of immune system class I antigen-presentation pathway, correlated to cognitive decline during ageing.

These is a mouse study. Now, this is a big caveat. In addition, though the authors believe they studied "cognitive" function in mice, the actual experiments, in my view, are based more on mouse locomotor skills.         

Initially, the authors reported that both in mice or humans ageing correlated with increased level of β2-microglobulin in the circulation.

Next, the authors showed that young 3 months old mice intravenously injected with β2-microglobulin made more errors in tasks requiring locomotor skills (water maze and fear freeze).

Interestingly, this negative effect of exogenous β2-microglobulin on mouse locomotive skills were short-lived (< 30 days, see d).

Importantly, β2-microglobulin-injected young TAP1 KO mice that lack surface MHC class I molecules did not show decline in their  locomotive skills.

Finally, old (17 months) β2-microglobulin KO mice showed less errors in their locomotive skills compared to their wild-type littermates.

In summary, this study claims that increase in β2-microglobulin level may indicate in locomotive skills in mice. Now, there are some major weakness in this paper that could affect its validity: (1) β2-microglobulin is a ~12kD size molecule that would have hard time to cross brain blood barrier (BBB); (2) there is substantial overlap between baseline and experimental errors in young mice in different figures (compare Fig. 1e vs. Fig. 2c vs. Fig. 4a); (3) outcome of experiments with β2-microglobulin injection into young β2-microglobulin KO are not done or not shown. 

David Usharauli 

Glycosylation directs IgG affinity selection during immune response

Immunoglobulin (Ig) Fc region determines Antibody effector functionality. In humans, for example, there are four types of IgG molecules that differ in their Fc domain structure: IgG₁, IgG₂, IgG₃, IgG₄. In addition, glycolysation (sugarization) of Fc fragments by syalic acid and fucose influences their binding to inhibitory or activatory Fc receptors.

In this new paper published in journal Cell, Jeffrey Ravetch's lab suggested a mechanism how glycolysation would optimize IgG affinity. This is a hybrid study involving both human and mouse studies. 

The authors have used influenza antigen HA as a model antigen. First, the authors showed that following flu vaccinination in healthy volunteers, overall abundance of sialylated IgG₁ Fc molecules (sFc) correlates with vaccine efficacy as measured by HAI.    

Next, to understand mechanism for improved HAI titers, the authors treated human B cells with immune complexes (ICs) derived either from sialylated (sIC) or asialylated (aIC) antigen-IgG conjugates. This experiments reveal that sialylated ICs induced up-regulation of inhibitory FcRIIb on antigen-specific B cells in a CD23-dependent manner. This could instruct antigen-specific B cells to undergo more stringent affinity-selection in germinal centers.

Antigen-antibody binding studies confirmed that in vivo immunization with sICs could elicit IgG₁ with higher affinity in a CD23-dependent manner (this is a mouse study).

Finally the authors showed that unlike pure HA antigen immunization, immune sera derived from mice immunized with sICs displayed superior activity against flu virus expressing stalk region (this region is a target for universal anti-flu antibodies).

In summary, these results suggest that efficacy of vaccination protocols is influenced by overall abundance of sialylated IgG₁ molecules. However, it is not clear whether this optimization of IgG₁ response are physiologically happening following primary (naive) or secondary (memory) immune response. Another point is whether this particular strategy with sICs is feasible for human vaccination protocols. 

David Usharauli