## ‘Mosquito smoothie’ innovation boosts future malaria vaccine potential

• June 17, 2021

A faster method for collecting pure malaria parasites from infected mosquitos could accelerate the development of new, more potent malaria vaccines.

The new method, developed by a team of researchers led by Imperial College London, enables more parasites to be isolated rapidly with fewer contaminants, which could simultaneously increase both the scalability and efficacy of malaria vaccines.

The parasite that causes malaria is becoming increasingly resistant to antimalarial drugs, with the mosquitoes that transmit the disease also increasingly resistant to pesticides. This has created an urgent need for new ways to fight malaria, which is the world’s third-most deadly disease in under-fives, with a child dying from malaria every two minutes.

Existing malaria vaccines that use whole parasites provide moderate protection against the disease. In these vaccines, the parasites are ‘attenuated’ – just like some flu vaccines and the MMR vaccine – so they infect people and raise a strong immune response that protects against malaria, but don’t cause disease themselves.

However, these vaccines require several doses, with each dose requiring potentially tens of thousands of parasites at an early stage of their development, known as sporozoites. Sporozoites are normally found in the salivary glands of mosquitoes, and in a natural infection are passed to humans when the mosquito bites. They then travel to the human liver, where they prepare to cause infection in the body.

Extracting sporozoites for use in a live vaccine currently requires manual dissection of the mosquito salivary glands – miniscule structures behind the mosquito head – by a skilled technician, which is a time-consuming and costly process.

The new method, described today in Life Science Alliance, vastly speeds up this process by effectively creating a ‘mosquito smoothie’ and then filtering the resulting liquid by size, density and electrical charge, leaving a pure sporozoite product suitable for vaccination. Importantly, no dissection is required.

Lead researcher Professor Jake Baum, from the Department of Life Sciences at Imperial, said: “Creating whole-parasites vaccines in large enough volumes and in a timely and cost-effective way has been a major roadblock for advancing malaria vaccinology, unless you can employ an army of skilled mosquito dissectors. Our new method presents a way to radically cheapen, speed up and improve vaccine production.”

But it’s not just about speed and cost. Traditional dissection methods struggle to remove all contaminants, such as proteins from the salivary glands, which are often extracted with sporozoites. The extra debris is likely to affect the infectivity of the sporozoites once they are inside the body, and could even affect how the immune system responds, impacting the efficacy of any whole parasite vaccine.

The new method also tackles this problem, resulting in pure uncontaminated sporozoite samples. The team discovered that, as well as being purer, sporozoites produced were surprisingly more infectious, hinting that vaccines produced using their method may require a much lower dose of sporozoites.

First author of the study Dr Joshua Blight, from the Department of Life Sciences at Imperial, said: “With this new approach we not only improve the scalability of vaccine production, but our isolated sporozoites may actually prove to be more potent as a vaccine, giving us additional bang per mosquito buck.”

The team developed and tested their method with both human and rodent malaria parasites. They then tested the rodent version as a vaccine in mice, and found that when exposed to an infected mosquito bite, vaccinated mice showed 60-70 per cent protection when immunisations were given into muscle. When the same sporozoites were given directly into the blood stream (intravenously) protection was 100 per cent, known as ‘sterile’ protection.

The researchers are now developing the method further in readiness for mass manufacture of sporozoites under good manufacturing practice (GMP) conditions in order to produce a vaccine ready for human challenge trials. The plan is that participants would be given vaccine-grade sporozoites produced using this method and then purposefully bitten by an infected mosquito.

Looking beyond vaccines the researchers also say their method should help accelerate studies of sporozoite biology in general, which could in turn lead to fresh insights into the liver stage of malaria and new drug and vaccine regimes.

###

The research was funded by the Wellcome Trust and the Bill & Melinda Gates Foundation.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.

## Third Dose of COVID Vaccine Boosts Protection in Transplant Recipients – Consumer Health News

• June 16, 2021

TUESDAY, June 15, 2021 (HealthDay News) — Researchers say an extra dose of two-dose COVID-19 vaccines may improve immune system protection for organ transplant patients, a group that’s so far responded poorly to two-dose vaccines.

“Our findings suggest clinical trials are warranted to determine if transplant recipients should receive COVID-19 vaccine booster doses as standard clinical practice, similar to what is currently done with hepatitis B and influenza vaccinations for this population,” said study lead author Dr. William Werbel. He is an infectious diseases research fellow at the Johns Hopkins School of Medicine in Baltimore.

People who receive a heart, lung, kidney or other solid organ transplant often take drugs to suppress their immune system and prevent rejection, but those drugs can interfere with the body’s ability to make antibodies in response to vaccines.

In two previous studies, only 17% of transplant recipients produced sufficient antibodies after one shot of a two-dose COVID-19 vaccine, and only 54% produced sufficient antibodies after the second dose, researchers reported.

Even transplant recipients who produced antibodies had levels well below those typically seen in people with healthy immune systems, the findings showed.

In the new study, the researchers evaluated 30 transplant recipients who previously received two doses of either the Moderna or Pfizer/BioNTech vaccine. None had reported an illness or a positive test for SARS-CoV-2 prior to vaccination. All were taking multiple immunosuppressive medications to prevent organ rejection.

Between March 20 and May 10, all participants got a third dose of either one of the Moderna or Pfizer vaccines, or they got the Johnson & Johnson shot.

“A third of the participants who had negative antibody levels and all who had low positive [antibody] levels before the booster increased their immune response after a third vaccine dose,” said study senior author Dr. Dorry Segev. He directs the Epidemiology Research Group in Organ Transplantation at Hopkins.

A week after receiving their third dose, 23 patients completed a questionnaire and some reported generally mild or moderate side effects. One patient had severe arm pain and another reported a severe headache. No patients reported fever or an allergic reaction.

There was one case of mild organ rejection, according to the report published online June 15 in the Annals of Internal Medicine.

Segev said the reactions seem acceptable, given the benefits that vaccines can confer.

Meanwhile, Werbel urged transplant patients and other immunocompromised patients to be careful.

“Although the third vaccine dose appears to raise the immune response of transplant recipients to higher levels than after one or two doses, these people may still be at greater risk for SARS-CoV-2 infection than the general population who have been vaccinated,” he said in a Hopkins news release.

“Therefore, we recommend that transplant recipients and other immunocompromised people continue to wear masks, maintain physical distancing and practice other COVID-19 safety measures,” Werbel added.

The American Society of Transplantation has more on COVID-19.

SOURCE: Johns Hopkins Medicine, news release, June 14, 2021

## Third Dose of COVID Vaccine Boosts Protection in Transplant Recipients | Health News

• June 15, 2021

By Robert Preidt, HealthDay Reporter

(HealthDay)

TUESDAY, June 15, 2021 (HealthDay News) — Researchers say an extra dose of two-dose COVID-19 vaccines may improve immune system protection for organ transplant patients, a group that’s so far responded poorly to two-dose vaccines.

“Our findings suggest clinical trials are warranted to determine if transplant recipients should receive COVID-19 vaccine booster doses as standard clinical practice, similar to what is currently done with hepatitis B and influenza vaccinations for this population,” said study lead author Dr. William Werbel. He is an infectious diseases research fellow at the Johns Hopkins School of Medicine in Baltimore.

People who receive a heart, lung, kidney or other solid organ transplant often take drugs to suppress their immune system and prevent rejection, but those drugs can interfere with the body’s ability to make antibodies in response to vaccines.

In two previous studies, only 17% of transplant recipients produced sufficient antibodies after one shot of a two-dose COVID-19 vaccine, and only 54% produced sufficient antibodies after the second dose, researchers reported.

Even transplant recipients who produced antibodies had levels well below those typically seen in people with healthy immune systems, the findings showed.

In the new study, the researchers evaluated 30 transplant recipients who previously received two doses of either the Moderna or Pfizer/BioNTech vaccine. None had reported an illness or a positive test for SARS-CoV-2 prior to vaccination. All were taking multiple immunosuppressive medications to prevent organ rejection.

Between March 20 and May 10, all participants got a third dose of either one of the Moderna or Pfizer vaccines, or they got the Johnson & Johnson shot.

“A third of the participants who had negative antibody levels and all who had low positive [antibody] levels before the booster increased their immune response after a third vaccine dose,” said study senior author Dr. Dorry Segev. He directs the Epidemiology Research Group in Organ Transplantation at Hopkins.

A week after receiving their third dose, 23 patients completed a questionnaire and some reported generally mild or moderate side effects. One patient had severe arm pain and another reported a severe headache. No patients reported fever or an allergic reaction.

There was one case of mild organ rejection, according to the report published online June 15 in the Annals of Internal Medicine.

Segev said the reactions seem acceptable, given the benefits that vaccines can confer.

Meanwhile, Werbel urged transplant patients and other immunocompromised patients to be careful.

“Although the third vaccine dose appears to raise the immune response of transplant recipients to higher levels than after one or two doses, these people may still be at greater risk for SARS-CoV-2 infection than the general population who have been vaccinated,” he said in a Hopkins news release.

“Therefore, we recommend that transplant recipients and other immunocompromised people continue to wear masks, maintain physical distancing and practice other COVID-19 safety measures,” Werbel added.

The American Society of Transplantation has more on COVID-19.

SOURCE: Johns Hopkins Medicine, news release, June 14, 2021

## Delaying a COVID vaccine’s second dose boosts immune response

• June 4, 2021

Facing a limited vaccine supply, the United Kingdom embarked on a bold public-health experiment at the end of 2020: delaying second doses of COVID-19 vaccines in a bid to maximize the number of people who would be at least partially protected from hospitalization and death.

Now, a study suggests that delaying the second dose of the Pfizer–BioNTech mRNA vaccine could boost antibody responses after the second inoculation more than threefold in those older than 801.

It is the first direct study of how such a delay affects coronavirus antibody levels, and could inform vaccine scheduling decisions in other countries, the authors say. “This study further supports a growing body of evidence that the approach taken in the UK for delaying that second dose has really paid off,” Gayatri Amirthalingam, an epidemiologist at Public Health England in London and a co-author of the preprint, said during a press briefing.

Many COVID-19 vaccines are given in two doses: the first initiates an immune response, and the second, ‘booster’ shot strengthens it. Clinical trials of the three vaccines used in the United Kingdom generally featured a three- to four-week gap between doses.

But for some existing vaccines, a longer wait between first and second doses yields a stronger immune response. Delaying the COVID-19 booster shots could also expand partial immunity among a greater swathe of the population than could the shorter dosing schedule. On 30 December, the United Kingdom announced that it would delay the second dose by up to 12 weeks after the first.

To determine whether the delay paid off, Amirthalingam and her colleagues studied 175 vaccine recipients older than 80 who received their second dose of the Pfizer vaccine either 3 weeks or 11–12 weeks after the first dose. The team measured recipients’ levels of antibodies against the SARS-CoV-2 spike protein and assessed how immune cells called T cells, which can help to maintain antibody levels over time, responded to vaccination.

Peak antibody levels were 3.5 times higher in those who waited 12 weeks for their booster shot than were those in people who waited only 3 weeks. Peak T-cell response was lower in those with the extended interval. But this did not cause antibody levels to decline more quickly over the nine weeks after the booster shot.

The results are reassuring, but are specific to the Pfizer vaccine, which is not available in many low-to-middle income countries, says Alejandro Cravioto, chair of the World Health Organization’s Strategic Advisory Group of Experts on Immunization. Countries will need to consider whether the variants that are circulating in their particular region might raise infection risk after only one vaccine dose, he says.

For the United Kingdom, extending the interval between doses was clearly the right choice, but the country’s lockdown deserves part of the credit for that success, says Stephen Griffin, a virologist at the University of Leeds, UK. “People are theoretically vulnerable between their first and second jab,” he says. “What’s worked in the UK is maintaining restrictions at the same time as vaccinating.”

## Delaying second Pfizer COVID-19 shot boosts immune response in over-80s, study finds

• June 3, 2021

Delaying the second dose of the Pfizer–BioNTech COVID-19 vaccine by 12 weeks after the first dose significantly boosts antibody responses in elderly people, according to a new U.K. study.

Researchers from the University of Birmingham and Public Health England (PHE) found that peak antibody levels were 3.5 times higher in those who waited 12 weeks for their booster shot, compared with those who had it after a three-week gap.

The study, which was published as a preprint and hasn’t yet been peer-reviewed, will lend support to the U.K. government’s decision at the start of its COVID-19 vaccine rollout in December 2020 to delay second doses of the vaccines to inoculate the elderly and vulnerable quicker, and ease the tight supplies.

At the time, the strategy divided experts, as drug regulators had authorized use of both the shot developed by drug company Pfizer
PFE,
+0.13%

and its partner BioNTech
BNTX,
-1.64%

and the shot produced by drug company AstraZeneca
AZN,
+0.20%

AZN,
-0.20%

with the University of Oxford on the basis of clinical trials that had spaced out the doses by only three or four weeks.

“This study further supports a growing body of evidence that the approach taken in the U.K. for delaying that second dose has really paid off,” said Dr. Gayatri Amirthalingam, a PHE epidemiologist and a co-author of the study, to reporters at a briefing.

Researchers said the study of 175 people aged 80 and over is the first direct comparison of how such a delay affects the immune response in any age group for the different intervals. An earlier study from the University of Oxford found that a single dose of the AstraZeneca–Oxford vaccine provides a high level of protection when boosters of the shot were delayed for 12 weeks.

The Birmingham researchers also looked at another part of the body’s immune response, found in T cells, which destroy any cells that have been infected with the virus. The peak T cell responses were higher in the group with a three-week interval between doses, the researchers found, but they cautioned that it wasn’t yet clear how protected individuals were based on which dosing schedule they received. “Research is required to further explore these variations in responses, the authors noted.

Separately on Friday, PHE said that COVID-19 vaccinations in the U.K, had directly prevented an estimated 11,700 deaths of people aged 60 and over by the end of April, and kept 33,000 people aged 65 and older out of hospital.

“As these figures highlight, getting your vaccine could save your life or stop you becoming seriously ill from COVID-19. It will also significantly reduce your chances of getting infected and infecting others,” said Dr. Mary Ramsay, PHE’s head of immunization.

“It is vital to get both doses of your vaccine when you are offered it,” she added.

## Millennial super mom Marian Rivera boosts her immune system with Calvit-C

• June 1, 2021
(The Philippine Star) – June 1, 2021 – 6:00am

MANILA, Philippines — Marian Rivera, millennial super mom and one of GMA Kapuso’s versatile actresses known for her roles in Marimar, Dyesebel, Amaya and Temptation of Wife, reveals how she keeps her immune system strong despite her busy lifestyle both as an actress and mom amid the pandemic.

Aside from working out regularly, she emphasizes the importance of taking vitamins that can help not only our bodies but also, and more importantly, our immune system.

Vitamins can naturally be obtained in healthy prepared meals. However, because of our busy schedules, we tend to neglect the importance of good nutrition and substitute it with high-cholesterol, sugary diets.

Vitamins help build and keep our body strong. Let us take for example these basic vitamins and their functions:

### 1. Vitamin A

This vitamin is key to good vision, a healthy immune system and cell growth. It is mostly found in eggs, broccoli, spinach and most dark green leafy veggies.

### 2. Vitamin B (B vitamins)

These vitamins have a direct impact on our body’s energy levels, proper brain function and cell metabolism.

They promote the growth of red blood cells, good digestion and proper nerve function. Examples of B Vitamin-rich foods are salmon, liver, milk, beef, oyster, clams and mussels.

### 3. Vitamin C

When we think of a vitamin that helps protect us from cough, colds and flu, one immediately comes to mind: Vitamin C.

Vitamin C, also known as Ascorbic acid is not only known for helping strengthen our immune system against harmful pathogens, but it is also responsible in keeping proper bodily functions via the development and repair of body tissues, formation of collagen and absorption of iron.

There are two types of vitamin C that are currently available in the market today—acidic and non-acidic. The market favors the non-acidic type for it is gentler on the stomach.

When it comes to non-acidic Vitamin C, Marian Rivera prefers Calvit-C.

Calvit-C is gentle on the stomach especially for those always on the go. It is made up of calcium and sodium ascorbate that may help protect the body against immune deficiencies, cardiovascular disease and help promote eye health.

Because it incorporates Calcium (in a form of calcium ascorbate) as one of its key ingredients, Calvit-C may also help in the development of strong bones and teeth, and promote healthy skin.

Sodium ascorbate, on the other hand, is a form of vitamin C that has sodium components, which makes absorbing Vitamin C easier, while lowering acidity levels of the body. Sodium ascorbate also serves as an antioxidant that helps keep cells healthy and protected from damage.

With benefits like these, Marian Rivera certainly has made a healthy choice in strengthening her immune system, the safe and non-acidic way.

• May 28, 2021

## INTRODUCTION

Nanoparticles (NPs) have attracted increasing attention in health care, biosensor, and immunotherapy research because of their excellent physicochemical properties (1, 2). NPs induce an immune system response once they enter or contact the body of humans or other organisms, and their pharmacokinetics also play critical roles in nanomedicine or other biologically related applications (35). For conventional biological experiments performed to assess the immune responses and organ burden of NPs, the cost is high, the reproducibility is low, the experimental time is long, and many animals are sacrificed (4, 6, 7). Quickly and accurately predicting the immune response and organ burden of NPs based on their physical and chemical properties is urgent for NP design and applications and pharmacokinetic assessments (8, 9). Given the complexity of the immune system and complicated properties of NPs, conventional methods [e.g., quantitative structure-activity relationships (QSARs) and molecular dynamics simulations] cannot precisely predict the immune response and organ burden of NPs.

The molecular structure must be known for QSARs and molecular dynamics simulations (10). However, the molecular structure of the immune system is difficult to depict. In addition, the complicated molecular calculations required for QSAR and molecular dynamics simulations require a long time to complete (11), which limits the ability to manage thousands of NPs with many properties (e.g., size, shape, and surface modifications) and the capacity to address data heterogeneity (12). As a type of robust nonparametric model, machine learning approaches, such as random forest (RF), artificial neural network (ANN), and support vector machine (SVM), have good potential to construct models that simulate complex relationships (12, 13). The anti-interference capability of machine learning may overcome data heterogeneity and provide a solution to predicting complicated biological responses to NPs (14, 15). However, a great challenge for machine learning methods is their poor interpretability, thus affecting the trustworthiness of models (16). Although high-precision predictions or classification tasks can be achieved by configuring the appropriate parameters (17, 18), the internal operations of the model are obscure, and poor interpretability also obscures causality (19). Current interpretable studies in the field of machine learning, such as prototype networks, local interpretable model-agnostic explanations, and Shapley additive explanations (2022), are devoted to revealing how machine learning works to achieve classification tasks to judge the rationality of decision-making, but they have not paid attention to the interaction among multiple features. Understanding the interaction among multiple features is useful to design NPs with ideal features and explore the mechanisms of bio-nano interactions.

To improve the interpretability of machine learning, it is urgent to understand how features affect labels and interact with each other (23), which widely occur in various machine learning models, such as RF and ANN (12, 19, 24). The present work proposed a feature interaction network concept along with a tree-based RF feature importance and feature interaction network analysis framework (TBRFA) as a proof-of-principle demonstration. TBRFA disassembles the trees implied by RF and then improves the interpretability of the RF model. The scheme is shown in Fig. 1. Inspired by meta-analysis workflow concepts and comparisons of RF, ANN, and SVM models, this study predicted the immune response and organ burden of various NPs with complicated properties. High correlation coefficients between the observations and predictions are achieved and further verified by validation set and animal experiments, thus ensuring the reliability of the model for the TBRFA framework. TBRFA includes two parts: importance analysis and feature interaction network analysis. TBRFA used a multiple-indicator importance analysis approach for RF based on existing methods to comprehensively screen the important features for the immune response and organ burden of NPs, which resolved the problems caused by the unbalanced data structure and the routine importance analysis method. Moreover, TBRFA proposes an interaction coefficient, uses the working mechanism of models to explore the interaction relationships among multiple features, and builds feature interaction networks to provide guidance for the design and application of ideal NPs.

## RESULTS

### Description and pretreatment of heterogeneous data

The data were extracted following the process in Materials and Methods. To comprehensively correlate immune responses and lung burden with the physical and chemical properties of NPs, data and literature on pulmonary immune responses and NP burden caused by lung exposure were assembled and integrated. On the basis of a comprehensive and rigorous extraction criterion of published data, a total of 1620 samples containing 16 features (including three parts: NP properties, animal properties, and experimental conditions) and 12 toxicity labels [biomarkers, e.g., total protein, total cells, and interleukin-1β (IL-1β)] were mined for immune response datasets. A total of 301 samples containing the same 16 features and 3 burden labels [i.e., lung, liver, and bronchoalveolar lavage fluid (BALF)] were mined for burden datasets. The references for the above samples are given in supplementary note S1. The datasets are described in Fig. 2 and tables S1 and S2. Six characteristic variables (NP type, shape, surface functionalization, rat/mouse, sex, and method) were described by the reported frequency (percentage). Inhalation as an exposure method was not recorded because of the difficulty in normalizing the inhaled doses. Ten numeric variables [diameter, thickness/length, zeta potential, specific surface area (SSA), mean age, mean weight, exposure duration, exposure frequency, recovery duration, and dose] were described by mathematical statistics (mean, SD, median, and distribution range). The abovementioned 6 characteristic variables and 10 numeric variables covered the main factors in the immune response analysis (25, 26). The distribution of the samples was visualized and arranged in descending order of NP length (Fig. 2). Heterogeneity of the immune response datasets was mainly caused by NP diversity (number of NP types, 57). The literature on biological responses usually focuses on widely used NPs [for example, multiwalled carbon nanotube (MWCNT) and TiO2 accounted for 22.54 and 16.6% in the immune response dataset, respectively], leading to a small proportion of some uncommon or previously unknown NPs. Encoding the discrete NP types into several continuous features reduced biases caused by the imbalance of NP types. The inconsistent NP characterization standards and differences in NP properties and the exposure protocols among laboratories also led to challenges for precise prediction. Systematic evaluations of the immune response to 57 NPs by routine experiments are costly and time-consuming, and it is also difficult to finish such complex model construction for QSAR analysis.

Many immune response biomarkers have been reported, although the choice of immune response index varies from laboratory to laboratory. Given the high reported frequency and the strong connections with immune responses (27, 28), the following 12 biomarkers were individually used for machine learning model construction: total protein, lactate dehydrogenase (LDH), total cells, macrophages, neutrophils), IL-1β, tumor necrosis factor–α (TNF-α), IL-6, IL-4, IL-10, monocyte chemoattractant protein-1 (MCP-1), and macrophage inflammatory protein–1α (MIP-1α). The burden datasets contained fewer NP types (n = 17) and had lower heterogeneity than the immune datasets. The NP burdens in the lung, liver, and BALF were chosen as labels. The details of the subsets are listed in table S3. General regression methods, such as multiple linear regressions, presented poor performance on heterogeneous data in complex systems because of the missing values (16). As shown in fig. S1, most of the multiple linear regressions obtained low correlation coefficients [coefficient of determination (R2), minimum value: 0.324] and high root mean square errors (RMSEs; maximum value: 0.902), especially in the large subsets, indicating that traditional regression methods were not suitable for heterogeneous data. In contrast, machine learning can achieve accurate predictions of these data, and the internal relationships of complex systems can be mined by TBRFA-based machine learning, as indicated in the following sections.

### Model performance and feature selection

Nonparametric and nonlinear machine learning methods have the ability to resist noise and are expected to build accurate prediction models using aggregated data (24). To eliminate the dimensional effects and balance the weights of features, z-score normalization and encoding of the character variables were applied before the model training (details are provided in Materials and Methods). The label values need to be normalized to improve the accuracy of the models. However, the distribution range of the label data was too wide (e.g., total proteins ranged from 1.5 to 2752.9%), and a considerable number of outliers occurred. The rough use of the z score can lead to serious collapse of the model accuracy. For the immune response dataset, we used the percentage change of the experimental group relative to the control group as the label and compressed the values between −1 and 1 (Materials and Methods, formula 2). For the organ burden dataset, we directly used the ratio of the detection concentration to the total exposure dose as the label. We initially compared the performance of ANN, SVM, and RF using all features for regression. To ensure that credible results are obtained from machine learning with a meta-analysis workflow, high R2 values are necessary (16). The R2 of the regression (Fig. 3, A to C) showed that in terms of the test set, the performance of RF (average of all models, 0.75 ± 0.12) was better than that of ANN (average of all models, 0.67 ± 0.11) and SVM (average of all models, 0.64 ± 0.10). Moreover, RF spent less time during the training process than ANN and had simpler adjustable parameters than ANN and SVM.

Figure S2 (A and B) shows that most of the features have low linear correlations, indicating that the features obtained via the literature and generated through coding will not cause overfitting due to multicollinearity. Figure S2 (A and B) also shows that there are low linear correlations between features and labels. The immune responses and accumulation burden of NPs are complicated because a single feature contributes little information to the label. Although the tested features may all affect the performance of models (29), a suitable feature selection procedure is still necessary to determine whether there is undetectable redundant information in the features. A greedy algorithm, sequential backward selection (SBS; see Materials and Methods), was used here to eliminate redundant information. Figure 3D shows that the SBS algorithm hardly improved the performance of the models, and we found that SBS abandoned some important NPs properties. Therefore, the RF models constructed on the basis of all features were selected for the subsequent analysis. The R2 values of the test set of most models were greater than 0.7, where R2 values for macrophages, lung burden, and BALF burden were >0.85, and the maximum value reached 0.896 (Fig. 3A and table S3). For some biological indicators, the testing R2 values less than 0.7 were probably due to the biases of data from interlaboratory studies. Moreover, we performed permutation tests, and the intercepts of the cross-validation coefficients (Q2) on the y axis were all less than 0.05, indicating that the models did not overfit (fig. S3) (30). Figure 3 (E to G) lists three examples (IL-6, IL-4, and lung burden) of regression results, and the others are given in fig. S4. To ensure that the features contributed valid information to the models, we performed feature value shuffling, and the predictive performance was abrogated after feature value shuffling (fig. S5). Because the published immune literature contained few negative samples (experimental group > control group), the scatters were mostly distributed in the top right of the scatter plots (Fig. 3, E and F), and the predicted results of the negative samples were distributed outside the RMSE interval. This flaw is inherently unavoidable because of the published datasets, although the prediction accuracy of the positive samples is not influenced. Multilabel prediction was performed by ANN, but the prediction accuracy rate was low, with most of the R2 values less than 0.7 (fig. S6), because the labels were independent of each other and represented different biological meanings.

NPs with a wide range of responses were used to verify the models. Five samples of each subset were randomly sampled as a validation set before building the model. The validation set did not participate in the construction of the model at all to ensure that the model did not learn them during the cross-validation. Figure 4A shows the prediction errors of the models on the validation set, where 76% of validation errors were less than 0.2. To verify the model further, animal experiments were performed in our laboratory. MWCNTs functionalized with triethoxycaprylylsilane and three MWCNTs of different sizes (small, medium, and large MWCNT, named S-MWCNT, M-MWCNT, and L-MWCNT, respectively) were chosen to conduct animal experiments. The above materials were all outside the scope of the datasets. Figure S7 shows the characterization of the NPs, and Fig. 4 (B to F) shows the immunofluorescence results of IL-1β. The expressions of IL-1β induced by MWCNTs distributed within the error bounds of the RF model (Fig. 4G). Figure 4G also contains the observation prediction of the validation set. The validation set of IL-1β contained five additional samples, which were randomly sampled from the main NPs in the IL-1 subset. The quantitative comparison between the observed and predicted results of the validation set indicated that the model was reliable (Fig. 4G). Compared with SVM and ANN, RF exhibited high accuracy and reliability for dealing with heterogeneous data in a complex system.

### Discovering important and unbiased features by TBRFA

The RF model was used to measure the importance of features by calculating the change in the error (increase in MSE) on the out-of-bag (OOB) data based on permutations of each feature (19). Figure 5 visualizes the feature importance measured by the MSE increase. Each model was normalized by its most important features. Figure 5A shows that the exposure dose (Fig. 5B) and recovery duration (the time from last exposure to animal euthanasia; Fig. 5C) have a great impact on the immune responses and NP burden. However, as a mathematical statistic, increases in the MSE are limited by the quality of the dataset. Thus, using increased MSE values as the only criterion for analyzing feature importance may lead to bias. For example, increased MSE values indicated that sex (male and female) contributed significantly to the IL-4 model (Fig. 5A). However, sex was not supposed to be the main factor affecting the immune response, and we confirmed this bias in the follow-up analysis. Therefore, absolute dominant features may be incorrectly identified when using a single importance evaluation index.

TBRFA uses multiple indicators to perform a multiway feature importance analysis and can evaluate the importance of features from different perspectives to balance the absolute dominant features achieved by a single indicator. Three other indicators, i.e., node purity increase, mean minimal depth, and P value, were combined with increased MSE values to comprehensively screen important features (Fig. 6, A and B). Although gender led to a higher increase in MSE and node purity than the other features, the P value indicated that its importance was not statistical significant (Fig. 6A). The distribution of the features’ mean minimal depth also suggested that the zeta potential, dose, length/weight, and SSA appeared more frequently near the roots than the other features (Fig. 6B). The TBRFA results of all models indicated that recovery duration and exposure dose were the main factors affecting the immune response and organ burden (Fig. 6, A and B, and figs. S8 to S21). TBRFA successfully overcame the importance bias caused by a single indicator with small datasets.

Figure 6C summarizes the conclusions of the TBRFA importance analysis. Exposure dose is well known as one of the most important features, while recovery duration, which represents the acute or chronic responses induced by NPs, has been ignored (9, 10). A partial dependence analysis (fig. S22A) showed that immune responses decreased as recovery duration increased and NPs tended to induce short-term acute immune responses. In the initial stage of exposure, the immune indicators increased rapidly; and within 50 days after the end of exposure, all immune indicators were decreased to less than 166.7% (corresponding to the normalized value 0.4 in fig. S22A) of the control group. The mechanisms of NP clearance by the immune system are complicated and include immune cell phagocytosis and extracellular traps (31). The clearance rate of NPs is also highly correlated with their structure and physical and chemical properties (32, 33). Figure S22B also confirmed that NPs could be gradually cleared by the immune system, with the accumulation of NPs in the lungs rapidly declining to approximately 20% of the total exposure dose within 100 days; hence, recovery duration became an important feature affecting the final observed immune responses. However, the immune responses did not decrease to the level of the control groups even after a long recovery duration. The optimized models proposed that the immune responses induced by NPs were persistent during the postexposure period; thus, the chronic and long-term toxicity of NPs deserve attention in future studies.

Although NP properties determine their immune response and organ burden, key properties of NPs remain inconclusive because of the limited number of animal experiments (34). The models proposed that SSA was important for immune responses, while diameter was important for organ burden compared with other properties (fig. S22, C and D). The SSA of NPs will allow the particles to adsorb proteins and form NP-protein corona complexes, which mediate immune responses (35). NPs with a large SSA are more likely to be internalized by the cell, thus causing a decrease in the number of molecules released by the cell (36). As the diameter decreased, fewer NPs stayed in the alveoli, and more NPs were transported to the lung and liver, which indicated that NPs with a small diameter (<100 nm) could easily penetrate biological membranes and achieve cross-organ transport (Fig. 6C and fig. S22D). A small size confers enhanced permeability to NPs, which is attributable to the binding to integrins on the surface of epithelial cells (37). Moreover, studies have shown that smaller NPs are less likely to be taken up by cells than larger NPs; thus, large NPs may preferentially translocate into the reticuloendothelial system, while small NPs may not be capable of being distributed among organs (38).

Animal studies, especially toxicological tests, are time-consuming, costly, and often impractical (10, 39). Because of these constraints, animal experiments can only screen the importance of specific or few properties (e.g., type, size, and surface functionalization) for one or a few NPs. In contrast, machine learning can quickly screen and sort various important features for various NPs at the same time. Moreover, the TBRFA approach overcame the conventional important feature analysis bias caused by the unbalanced data structure, identified the critical features (recovery duration, dose, SSA, and diameter), and quantified the importance of features and the response degree of biological biomarkers to features, thus providing a reference for the design of ideal NPs.

### Feature interaction networks created by TBRFA

The mechanism underlying the ability of NPs to induce immune toxicity is complicated. Understanding how the different properties of materials interact with each other (feature interaction networks) and influence their immune response and organ burden is critical to the design and application of NPs, although the related information remains largely unknown (39). Because most machine learning methods are black-box models, identifying the interactions among features represents a difficult challenge. The conditional minimal depth (Fig. 7A) in RF represents the strength of interactions between two features (40). As demonstrated in Fig. 7B and figs. S8 to S21C, the conditional minimal depth was calculated to obtain the order of the interaction strength between features, and then the strongest four feature interaction relationships were displayed through a double-variable partial dependence analysis (Fig. 7, C to F, and figs. S8 to S21, D to G). However, strong interactions between features were not found for most of the partial dependence analyses, and they tended to show a simple additive effect. A comparison with the unconditional minimal depth of the feature showed that the conditional minimal depth was greatly affected by the importance of the feature itself. If the two features are both important, then they will have small and similar depths; therefore, even if they are close in the tree, they may not interact with each other.

To weaken the influence of feature importance on the interaction represented by the conditional minimal depth, the RF based on the decision tree was explored again, which was called feature interaction network analysis. We redesigned an interaction coefficient on the basis of the conditional minimal depth (see Materials and Methods and Fig. 8A). To calculate the interaction coefficient between features, we explored the working mechanism of RF in depth, traversed each decision tree constituting the RF, and obtained the properties of the tree and its features. Subsequently, we integrated the interaction coefficients and built interaction networks for each RF model. The interactions between NP properties were mainly analyzed to provide guidance for NP design. Interaction networks of total proteins, total cells, IL-6, IL-4, lung burden, and liver burden are provided as examples in Fig. 8 (B to G), and the other networks are given in figs. S22 to S30. As shown in Fig. 8 (B to G), the network analysis for the immune response dataset indicated that SSA had a strong interaction with the zeta potential and length/width, while the network analysis of the burden dataset indicated that diameter had strong interactions with length. The above feature interaction network explained the critical roles of SSA in immune responses (fig. S22C). The feature interaction network clearly showed the connections between NP properties and immune responses or organ burden (Fig. 9, A to F). For example, NPs with negative charges (<0 mV) and small SSAs (0 to 200 m2/g) induced low levels of total protein and total cells increase and high levels of IL-4 and IL-6 release (Fig. 9, A to D). The total protein value reflects the trend of inflammation as a whole, and the complicated trend between different cytokines and NP zeta potentials may be related to the electrostatic interaction between the particle surface and cytokines (41, 42). There was an important influence of both zeta potential and SSA on the formation of protein coronas and the uptake of nanomaterials (43, 44). The interactions excavated by TBRFA indicated that both factors were mutually restrictive and affected the biocompatibility and toxicity of NPs. Long NPs caused more severe immune responses than short NPs, while a suitable SSA reduced different immune responses (Fig. 9, E to F). For the burden dataset, the interactions between diameter and length caused the opposite degree of lung and liver burden. One-dimensional NPs had the lowest accumulation in the lung and the highest accumulation in the liver, suggesting that length determined the translocation capacity of NPs, which was consistent with the conclusion drawn by Huang et al. (45) that long NPs were cleared from the lung quickly. The above results also agreed that gold nanorods accumulated significantly less than gold spheres with similar size and surface chemistry (46). NPs with nonspherical shapes exhibited the properties of targeted accumulation because of the deviating hydrodynamic behavior (e.g., roll and rotate) and lateral drifting speed in the vessel (47). In addition, NPs with a length and width in the range of 100 and 500 nm showed greater transport across organs (Fig. 9, G to I). Figure 9J visualizes the conclusions of the above partial dependence analysis. This result further supplements the conclusion on the NP size effect on organ burden obtained from the TBRFA importance analysis, indicating that feature interaction network analysis was a powerful method of boosting the interpretability of machine learning. The models are consistent with the experiments, ensuring the reliability of TBRFA. TBRFA discovers hidden feature interactions that are difficult to explore through few experiments and other machine learning methods [e.g., ANN and deep neural network (DNN)]. TBRFA extracts the complex interaction network among NP properties, immune response, and organ burden, thereby providing guidance for the design and discovery of ideal NPs.

## DISCUSSION

Given the high costs and limitations associated with animal protection, comprehensive biological response evaluation experiments for various NPs are not practicable. Although a number of studies have applied machine learning approaches to solve the above problems, the interpretability of the models is poor, thus hindering the application of machine learning in the design and discovery of ideal NPs. Here, a rigorous TBRFA approach that boosts the interpretability of machine learning successfully predicted the pulmonary immune response and organ burden of NPs. The optimized prediction accuracy was achieved, with R2 values of all training sets >0.9 and half of the test sets >0.75. To overcome the shortcomings of traditional importance analyses, TBRFA used multiway importance analysis and reduced the biases caused by the unbalanced structure of small sample datasets. The TBRFA framework also established feature interaction networks and boosted the interpretability of machine learning. It is difficult for researchers to fully explore the joint effects of multiple features through experiments. In most machine learning studies, the exploration of interpretability usually stops at revealing the importance of features and ignores the relationships under the joint action (e.g., antagonism and synergy) of multiple features (12, 48). It is well known that the interactions between NPs and biology are determined by multiple features or properties of NPs (49), but the key features that regulate biological responses remain controversial. For the design of NPs, studies usually focus on one or two features (50), but how the other features influence the global properties of NPs is unclear. The feature interaction network analysis in Fig. 8 provides insights into the above challenging question.

The TBRFA approach revealed and explained the critical roles of SSA and diameter in the NP immune response and organ burden, respectively. The improved interpretability of machine learning is useful for models that explore causation and the design of excellent NPs in the medicine, biosensor, drug delivery, and other health care fields, and it also provides accurate predictions. Some animal experiments conducted by different laboratories are not completely consistent due to the biological response affected by multiple factors (e.g., the sources of animals, the environments of animal growth, and the operation skills of researchers) (51, 52). The above biases of data, especially for small-size data, would lead to a rather large variation in test predictive performance across labels. As the above explanation, the biases of data from interlaboratory studies may contribute to the misclassified (very off-diagonal) samples in Fig. 3E. Drafting standardized nanomaterial characterization protocols and animal exposure protocols and improving the quality of the literature will ensure the authenticity of the label value and greatly improve the accuracy and applicability of TBRFA (49). Further improving the degree of automation of interaction coefficient calculations will also make TBRFA more suitable for big data. Overall, the models established in this work are suitable for predicting the lung immune response induced by inorganic NPs and their lung burden. TBRFA also deserves application in other fields, and the discovered feature interaction networks could contribute to human disease or anticancer drug research.

## MATERIALS AND METHODS

### Extracting data to establish datasets

The immune response data were obtained from published articles in the ISI Web of Knowledge database (the datasets before 31 December 2020 were collected). A total of 2548 studies were initially searched using the following search formula: TS = nano* AND TS = immun* NOT TS = Immunosensor AND (TS = mice OR TS = rat OR TS = mouse) AND (TS = pulmonary OR TS = lung). Given the heterogeneity of the biological data, the acquired publications were then filtered by the following conditions: (i) full text was available; (ii) the topic was the pulmonary immune responses of mice or rats induced by NPs; (iii) at least one of the following immune metrics contained: total protein, LDH, alkaline phosphatase, cell count [total cells, macrophages, and neutrophils (polymorphonuclear neutrophils)], cytokines, and chemokines; (iv) exposure methods were instillation, oropharyngeal aspiration, or similar methods; and (v) basic material characterization data and experimental conditions were provided. Last, 1620 samples were identified to establish datasets. To precisely and comprehensively establish the relationships between NP/exposure features and pulmonary immune responses, 16 features were included in the dataset, and they consisted of seven material properties, four animal properties, and five experimental conditions. Two biochemical indicators, three types of cell counts, and six different cytokines were selected as labels for the immune responses. The details are shown in supplementary Excel files.

Acquisition of lung burden data followed the above workflow. A total of 3525 studies (search date, June 2020) were initially searched using the following search formula: TS = nano* AND (TS = mice OR TS = rat OR TS = mouse) AND (TS = pulmonary OR TS = lung) AND TS = (accumulat* or burden* or clear*). Taking the exposure method and the determined features from the immune response dataset as the main filtering basis, 302 samples were lastly used to establish datasets. Liver and BALF burden data were also included in this burden dataset (supplementary Excel files). The data of biological responses to NPs are very complex and distributed in the texts, tables, and figures of publications. It is difficult to extract the required data by machine. The data were extracted from the publications by hand. The data directly given in texts and tables were copied by hand. For the data in the figures, the “Digitizer” tool provided by OriginLab was used to read each point three times and then the averages were calculated. To reduce the errors, the extracted data were checked thoroughly by another person again.

### Preprocessing data

To calculate characteristic variables, the six characteristic variables in the dataset (NP type, shape, surface functionalization, animal, gender, and method) need to be coded. “One-hot” is a general encoding method that converts disordered and discrete variables into binary vectors to overcome the problem that such variables cannot be recognized by machine learning algorithms. For the three discrete features with few unique types (i.e., animal, gender, and method), the “one-hot” coding was adopted. For one-hot encoding features, there were no obvious correlations through correlation coefficient analysis (fig. S2). However, the one-hot method was not suitable for the descriptor “NP types” because the diversity of NP types (n = 57) caused a rapid increase in the data dimension. New descriptors were created on the basis of chemical properties for the other three characteristic variables to distinguish different properties. The new descriptors are as follows:

1) NP types: carbide (0/1), macromolecular compound (M.C., 0/1), oxide (0/1), salt (0/1), component 1 (Com.1, relative atomic mass), and component 2 (Com.2, relative atomic mass).

2) Shape: hollow (0/1), granular NPs (Dim.0, 0/1), one-dimensional NPs (Dim.1, 0/1), and two-dimensional NPs (Dim.2, 0/1).

3) Surface functionalization: positive charge (S.P., 0/1) and negative charge (S.N., 0/1).

To prevent certain features from contributing excessively to the models, the z-score (Eq. 1) normalization method was applied to the features. The formula is listed as follows

$x′=(x−μ)σ$

(1)where x is the feature value, μ is the mean of each feature, and σ is the SD of each feature. The raw data in literatures are difficult to use for direct comparison. The difference in experimental design and used animals may produce biases in models. To intuitively reflect the relative degrees of the immune response caused by different NPs in different literatures, the label data were converted to values between −1 and 1 on the basis of the treated and control groups in each literature individually, rather than using the data in all literatures together, to reduce the biases.

The formula is listed as follows

(2)where y and c are the values of the experimental group and control group, respectively.

### Machine learning regression

RF models were trained using scikit-learn in Python 3.7. According to different labels, the datasets were split into 15 subsets (12 immune response subsets and 3 burden subsets) that corresponded to 15 regression models. The RF models used 500 random decision trees and selected 5 random features at each node, which were determined by the grid search method (53). RF is based on bootstrapping to avoid inherent overfitting (54, 55). A 10-fold cross-validation (ShuffleSplit) method was used for each machine learning algorithm to prevent overfitting, with 90% of the samples in each subset chosen as training sets and 10% chosen as test sets. Approximately 36.8% of the samples in the training set, which were called OOB data, were used as the validation sets and not used during the training process (56). The percentage of OOB was calculated by the following formula

$limm→∞(1−1m)m=1e≈0.368$

(3)where m is the frequency of sampling and e is the natural constant with a value of approximately 2.71828. Each model was trained 10 times, and the average of the R2 and RMSE values between the predictions and observations were calculated to measure the model performance.

Two-layer fully connected ANN models were trained using Keras in Python 3.7. The workflow was similar to that of the RF regression except that 25% of the samples in the training set were split randomly as validation data to monitor overfitting. Multilayer feedforward neural networks with different layers and different units in the hidden layer were trained to find the optimal configuration, and 57 and 55 units in the hidden layers were adopted in the ANN immunotoxicity and burden models, respectively. Stochastic gradient descent was used as the optimizer, and a small learning rate decay (less than 0.0001) was set to prevent overfitting during the training process. SVM models were trained using scikit-learn in Python 3.7. Tenfold ShuffleSplit method was also applied. “Rbf” was chosen as the kernel function. The regularization parameter was set to 1.

### Overfitting test

To judge whether the RF models used were overfitted, we adopted a permutation test method. In each process of 10-fold cross-validation, 20, 40, 60, 80, and 100% of the label values in the training set were randomly replaced by random values within the original label range, and the corresponding cross-validation coefficients (Q2) were calculated. The calculation formula of Q2 is as follows

$Q2=1−∑i=1n(yi−ŷ)2∑i=1n(yi−y¯)2$

(4)where yi is the observed label value,

$ŷ$

is the predicted label value, and

$y¯$

is the average of the label. The permutation of each ratio (20, 40, 60, 80, and 100%) was performed 10 times, resulting in 500 permutation Q2 values for each model (5 ratio × 10 times/ratio × 10 folds). Subsequently, linear regressions were performed on the Q2 values and the correlation coefficients between the original labels and the permutation labels. The intercept of the regression result on the y axis less than 0.05 proves that the model is not overfitting (30).

### NP preparation and characterization for model verification

Three types of MWCNTs with axial lengths ranging from approximately 0.5 to 2 μm were obtained from XFNANO (China): S-MWCNTs (production number XFM10; diameter, 8 to 15 nm; purity > 95%), M-MWCNTs (production number XFM16; diameter, 10 to 20 nm; purity > 95%), and L-MWCNTs (production number XFM28; diameter, 30 to 50 nm; purity > 95%). All the MWCNTs were produced by the chemical vapor deposition method. To investigate the effects of surface modification on immune responses, nano-TiO2 (production number XFI02; diameter, 20 to 40 nm; purity > 99%) and hydroxylated MWCNTs (production number XFM16; diameter, 10 to 20 nm; purity > 95%) functionalized with triethoxycaprylylsilane (F-MWCNT) were synthesized. Functionalization was carried out by stirring a mixture of 100 mg of hydroxylated MWCNTs with 10 mg of triethoxycaprylylsilane (DB, Shanghai) in 100 ml of 90% ethanol solution at 70°C for 8 hours. The mixture was then centrifuged at 1699g for 15 min, washed with Milli-Q water (18.2 megaohm cm−1), and then vacuum freeze-dried. To calculate the sizes of the NPs, transmission electron microscopy (TEM) images were obtained on a TEM instrument (JEM-2010 FEF, JEOL, Japan). Fourier transform infrared (FTIR) spectroscopy (Bruker Tensor 27, Germany) with a resolution of 2 cm−1 from 4000 to 400 cm−1 was used to confirm the synthesis of F-MWCNTs. Zeta potentials were determined by a ZetaSizer Nano instrument (BI-200SM, Brookhaven, USA). BET surface areas were measured with a surface area and porosity analyzer (ASAP 2460, Macrometrics, USA). The results of characterization were given in the supplementary note S2 and tables S4 and S5.

### Animal experiment

Male Institute of Cancer Research (ICR) mice with a body weight range of 24 to 26 g were purchased from Vital River Laboratories (Beijing, China) and acclimated for 2 weeks in animal facilities supplied with normal food and water. All animal studies were performed in accordance with the guidelines and regulations of the Human and Animal Experiments Ethical Committee of the Nankai University. The four types of NPs were individually suspended in sterile phosphate-buffered saline (PBS) at 1 mg/ml. Suspensions were processed with ice-bath ultrasonication before instillation to ensure even dispersion. ICR mice (n = 6 per group) were treated individually with a single intratracheal instillation of 25 μl of (1 mg/kg) NPs. Equal volumes of PBS were instilled as a control.

### Immune response analysis

Frozen sections (thickness, 5 μm) of lung lobes were made by a clinical cryostat (CM1850, Leica, Germany) for immunofluorescence. Anti–nuclear factor κB (NF-κB) p65 with Alexa Fluor 555 (bs-0465R-AF555, Bioss, China) and anti–IL-1β with fluorescein isothiocyanate (FITC) (bs-0812R-FITC, Bioss, China) were used to label NF-κB p65 and IL-1β, respectively. 4′,6-diamidino-2-phenylindole (DAPI) was used to stain the nuclei throughout the lung section. Immunofluorescence images of NF-κB (red), IL-1β (green), and DAPI (blue) were obtained with a confocal laser scanning microscope (LSM880 with Airyscan, Zeiss, Germany) at 543, 488, and 405 nm, respectively.

### Model validation

Model validation was composed of two parts: animal experimental and validation sets. In terms of animal experiments, the characterization data and experimental conditions of NPs were input into the trained RF models to obtain predicted values. To validate the accuracy of the model, the immunofluorescence intensity of the IL-1β protein was statistically analyzed and compared with the predicted IL-1β results. The interval formed by ±RMSE was considered to be the allowable error bound of the model (11). As the validation set, random seeds were set to randomly sample the validation set (n = 5) for each subset before modeling. The five validation sets covered the NPs with a wide range of responses. The validation set did not participate in the construction of the model at all to ensure that the model did not learn them during the cross-validation. During the 10-fold cross-validation process, the average predicted values of 10 times were obtained for comparison with the observed values. According to the data source, the animal experiment validated the three critical NP properties (diameters, SSA, and zeta; see table S4), and the validation set validated all features.

### Feature selection

Models were optimized using a SBS procedure. This method started from the full set of features to eliminate redundant features one by one to find the optimal feature subset. The R2 of each selection process was calculated and recoded to compare the performances of different feature combinations. To avoid the models losing too much information, we set the minimum number of features selected by SBS to not be less than one-half of the total number of features.

### Feature importance analysis of TBRFA

Feature importance analysis was based on the “randomForest” and “randomForestExplainer” packages in R 4.0.2. To avoid the bias produced by a single indicator (i.e., MSE) used in comment RF, a total of four indicators (MSE increase, node purity increase, P value, and mean minimal depth) were selected to represent different perspectives and to comprehensively evaluate the importance of features. MSE increase is based on the decrease in predictive accuracy of the forest after perturbation of the variable; node purity increase is based on changes in node purity after splits on the variable; and P value is based on the one-sided binomial test to evaluate the significance of feature importance (40, 53). The mean minimal depth is based on the structure of the forest. The significant and important features were screened by calculating the MSE and node purity (measured by residual sum of squares) increase and the P value. During the training process, feature importance was also reflected by the mean minimal depth of the feature among trees. In this study, five random features were set to be randomly sampled as candidates at each node and one of the five that contributed the most to the overall split was retained at the node. Therefore, features near the root were more important than others.

### Feature interaction network analysis of TBRFA

The conditional minimal depth of features represents the strength of the interaction between two features. The original conditional minimal depth of features was initially calculated using the randomForestExplainer package. Four groups of strong interaction features were selected for the double-variable partial dependence analysis. However, there was actually no specific interaction between some features. To improve the present method and weaken the influence of the feature importance (minimal depth) on the interaction among features, we explored the conditional minimal depth of each feature in each tree of an RF model and used the occurrences of interaction among the feature and its corresponding root feature to normalize its conditional minimal depth. The interaction coefficient is defined as follows

$ρA:B=OA:BOB×∑i=1ntreeDT−DA:B−1DTOA:B$

(5)where A and B are two of the features, A:B means the interaction of B and the maximal A-subtree [see the description by Ishwaran et al. (40)], ρA:B is the directed interaction coefficient, OA:B is the occurrence of the interaction of B and the maximal A-subtree among the trees (ntree = 500), OB is the occurrence of B among the trees (ntree = 500), DT is the tree depth, and DA:B is the minimal depth of B in the maximal A-subtree, that is, the conditional minimal depth. All interaction coefficients in each RF model were calculated, and the interaction coefficients that had two directions (using average values) were merged. Then, the importance of features was combined to build feature interaction networks, and new strong interaction features for double-variable partial dependence analysis were screened out.

### Statistical analysis

A two-sided Kolmogorov-Smirnov test was implemented in R software to test whether the distribution of the data was normal (P > 0.05).

Acknowledgments: We acknowledge the technical support provided by H. Zhang, C. Zou, J. Xu, W. Li, L. Zhang, and R. Cai from the Nankai University. Funding: This work was financially supported by the National Natural Science Foundation of China (grant nos. 21722703 and 42077366), the 111 program (grant no. T2017002), and the National Key Research and Development Project (grant no. 2019YFC1804603). Author contributions: X.H. designed the project; X.H. and F.Y. contributed to the writing and revision of the manuscript; F.Y. ran the models; F.Y. and P.D. collected the data; C.W., F.Y., and T.P. performed the animal experiments. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Datasets containing all relevant literature data are provided as supplementary Excel files, and the datasets are also available on Zenodo at https://doi.org/10.5281/zenodo.4661099. Performance of models, TEM, FTIR, and other characterization of materials are also provided in the Supplementary Materials. The custom codes written to develop the machine learning models, perform TBRFA and other analysis, and generate figures are all available on Zenodo at https://doi.org/10.5281/zenodo.4660115.

## Breast milk boosts immunity against infections: Top pediatrician

• May 23, 2021

As we spend more and more time living in pandemic conditions the state and importance of our health come ever to the forefront and the health of our children most of all. So, parents are looking at ways on how they can maximize the protection of their babies. According to a top Turkish pediatrician, there is a simple answer: breast milk.

Breast milk, not only provides nutrition to infants but it strengthens the immune system and protects them from infections. “Studies show that infants who are breastfed for a year are 50% less likely to catch an infection than other babies,” said Nalan Karabayır, who teaches pediatric health and diseases at Medipol Mega University Hospital in Istanbul.

Breast milk has indirect ways of helping infants as well. Antibodies for the coronavirus were found in the milk of mothers who were vaccinated, and babies can be protected in this way.

Karabayır said if the mother is inoculated against the coronavirus, it will protect her and her baby from infection.

Touching on the effects of COVID-19 on mothers and infants, and referring to studies, Karabayir noted that the pandemic affects children of all ages but with our current knowledge, it is accepted that it is not transmitted through breast milk.

“It is known that antibodies against SARS-CoV-2, which occur in the mother who had a COVID-19 infection, also pass into breast milk and protect the baby from infection,” she said.

It is known that breastfeeding is safe as long as mothers strictly follow isolation practices such as wearing masks, social distancing and hygiene, she said.

Underlining that living cells of the mother’s immune system are present in breast milk, Karabayır explained, “Thanks to breast milk, the baby receives 1.5 million live cells in every 1 milliliter (0.03 ounce) of milk.”

The World Health Organization and groups dealing with infant and children’s health suggest that babies should breastfeed for the first six months and breastfeeding should continue until at least they reach 2 years old, she said.

Karabayır emphasized that babies are born before their immune systems are fully mature. “Physical and chemical conservators are not yet developed postnatally. For this reason, they need immunological components in breast milk to combat microorganisms that may cause infection.”

“In addition to the nutritional properties of breast milk, the live cells, probiotics, cytokines, immunoglobulins and oligosaccharides it contains, provide protection of the baby from infections. For these reasons, breast milk is unique,” she added.

## Novel immunotherapy boosts long-term stroke recovery in mice

• May 19, 2021

PITTSBURGH, May 19, 2021 – Specialized immune cells that accumulate in the brain in the days and weeks after a stroke promote neural functions in mice, pointing to a potential immunotherapy that may boost recovery after the acute injury is over, University of Pittsburgh School of Medicine neurologists found.

The study, published today in the journal Immunity, demonstrated that a population of specialized immune cells, called regulatory T (Treg) cells, serve as tissue repair engineers to promote functional recovery after stroke. Boosting Treg cells using an antibody complex treatment, originally designed as a therapy after transplantation and for diabetes, improved behavioral and cognitive functions for weeks after a stroke in mice compared to those that did not receive the antibody complex.

“The beauty of this treatment is in its wide therapeutic window,” said senior author Xiaoming Hu, M.D., Ph.D., associate professor in the Department of Neurology at Pitt’s School of Medicine. “With most strokes, you have four and a half hours or less when you can give medication called tPA to reopen a blocked blood vessel and expect to rescue neurons. We’re excited to identify a mechanism that may promote brain recovery by targeting non-neuronal cells well after this window closes.”

Previous research in stroke has been focused on developing new drugs to reduce neuronal death. And whereas these acute stroke treatments quickly lose effectiveness after neurons die, Treg cells remain active for weeks after the injury.

True to their name, Treg cells are immune cells that regulate the immune response, including curtailing excessive inflammation that could harm more than help. Hu and her colleagues observed that the levels of Treg cells infiltrating the brain began to increase about a week after a stroke and continued increasing up to five weeks later. So, they did multiple tests in mice after they’d had strokes, paying particular attention to the brain’s white matter–which is the brain tissue through which neurons pass messages, turning thoughts into actions, like lifting food to your mouth or saying the name of an object you’re looking at.

Mice that were genetically unable to produce Treg cells fared worse than mice with a robust Treg cell response. Interestingly, it was only in the latter phases of stroke recovery that the Treg cell-depleted mice suffered impairments in white matter integrity and behavioral performance compared to mice with a normal Treg cell response.

Additionally, when normal mice were given an antibody complex called “IL-2:IL-2Ab” to boost their Treg cell levels after a stroke, their white matter integrity improved even more and neurological functions were rescued over the long term. The mice with more Treg cells had an easier time moving and had better memories, allowing them to navigate mazes faster after a stroke than their non-treated counterparts.

“This strongly suggests that, rather than working to preserve white matter structure and function immediately after a stroke, Treg cells influence long-term white matter repair and regeneration,” said Hu, also a member of the Pittsburgh Institute of Brain Disorders and Recovery and a U.S. Department of Veterans Affairs (VA) investigator. “Our findings pave the way for a therapeutic approach to stroke and other neurological disorders that involve excessive brain inflammation and damage to the white matter. Treg cells appear to hold neurorestorative potential for stroke recovery.”

Hu stressed that there are still many hurdles to cross before Treg cells could be used in humans for stroke recovery. Namely, research is needed to determine the best way to boost the number of Treg cells in stroke victims. This could be done by improving IL-2:IL-2Ab so that it better stimulates production of Treg cells with fewer side effects, or a personalized therapy could be developed where some cells are taken from an established Treg cell bank and used to grow custom Treg cells in the lab, which could then be infused back into the patient.

###

Additional authors on this research are Ligen Shi, M.D., Ph.D., Zeyu Sun, M.D., Wei Su, M.D., Di Xie, M.D., Qingxiu Zhang, M.D., Xuejiao Dai, M.D., Ph.D., Kartik Iyer, B.S., T. Kevin Hitchens, Ph.D., Lesley M. Foley, B.S., Sicheng Li, Ph.D., Donna B. Stolz, Ph.D., Kong Chen, Ph.D., Ying Ding, Ph.D., and Angus W. Thomson, Ph.D., all of Pitt; Fei Xu, B.S., and Jun Chen, M.D., both of Pitt and the VA Pittsburgh Health Care System, and Rehana K. Leak, Ph.D., of Pitt and Duquesne University.

This research was supported by the National Institute of Neurological Disorders and Stroke grant NS094573 and VA grant I01 BX003651.

To read this release online or share it, visit https://www.upmc.com/media/news/051921-hu-treg-stroke.

About the University of Pittsburgh School of Medicine

As one of the nation’s leading academic centers for biomedical research, the University of Pittsburgh School of Medicine integrates advanced technology with basic science across a broad range of disciplines in a continuous quest to harness the power of new knowledge and improve the human condition. Driven mainly by the School of Medicine and its affiliates, Pitt has ranked among the top 10 recipients of funding from the National Institutes of Health since 1998. In rankings recently released by the National Science Foundation, Pitt ranked fifth among all American universities in total federal science and engineering research and development support.

Likewise, the School of Medicine is equally committed to advancing the quality and strength of its medical and graduate education programs, for which it is recognized as an innovative leader, and to training highly skilled, compassionate clinicians and creative scientists well-equipped to engage in world-class research. The School of Medicine is the academic partner of UPMC, which has collaborated with the University to raise the standard of medical excellence in Pittsburgh and to position health care as a driving force behind the region’s economy. For more information about the School of Medicine, see http://www.medschool.pitt.edu.

###

Contact: Allison Hydzik

Office: 412-647-9975

Mobile: 412-559-2431

E-mail: HydzikAM@upmc.edu

Contact: Anastasia Gorelova

Mobile: 412-491-9411

E-mail: GorelovaA@upmc.edu

## Immunotherapy Approach against Hepatitis B Virus Boosts T Cells and Acts as Direct Antiviral

• May 17, 2021

Scientists at University College London (UCL) have identified a new immunotherapy against hepatitis B virus (HBV), the world’s most common cause of liver cancer. The team’s study using immune cells isolated directly from patient liver and tumor tissue, showed that using a known drug compound to block the activity of an enzyme called acyl-CoA:cholesterol acyltransferase (ACAT) was highly effective at boosting the response of specific immune system T cells that can fight both the virus and associated tumors. ACAT is required for cholesterol esterification, a mechanism that prevents the accumulation of excessive, potentially toxic levels of cholesterol in cells. The study, in addition, showed that ACAT inhibition acted as an antiviral that directly reduced HBV replication.

Commenting on the findings, Nathalie Schmidt, PhD, from UCL’s division of infection & immunity, said, “We have found a highly effective novel target for the treatment of chronic hepatitis B virus infection and liver cancer. Modulating cholesterol metabolism with ACAT inhibitors has the unique features of directly targeting the virus and tumors while at the same time boosting the T cells that fight them. This enables us to tackle the disease from multiple directions at the same time.”

The results are reported in Nature Communications by the UCL research team and collaborators at Oxford University, the Royal Free London NHS Foundation, and Leiden University Medical Centre. Schmidt is first author of their published paper, which is titled, “Targeting human Acyl-CoA:cholesterol acyltransferase as a dual viral and T cell metabolic checkpoint.”

Each year, chronic HBV (CHB) infection causes an estimated 880,000 deaths globally from liver cirrhosis and liver cancer (hepatocellular carcinoma; HCC), the authors explained. “Worldwide, HBV is the commonest cause of HCC, and HCC is the third most common cause of cancer-related deaths …” Current treatments for chronic HBV infection require long-term antiviral suppression using drugs known as nucleos(t)ide analogs (NUCs), but such therapies don’t completely inhibit ongoing expression of viral antigens, and can cause off-treatment effects.

Immune cells such as T cells are indispensable for fighting viruses and tumors, the UCL team noted, but are often highly dysfunctional and fail to control these diseases. Current standard-of-care treatments for HBV are thus largely incapable of eliminating the virus, do not prevent cancer development, and do not rescue immune cells. And while immune checkpoint inhibition using PD-1 blockade has been tested in Phase III HCC trials and in a Phase Ib study in HBV patients without HCC, the authors noted, “ … in both settings only a minority of patients had sustained responses, underscoring the need for additional approaches to increase therapeutic effect.” They noted, “A pressing goal is to develop a combination of more potent antiviral and immunomodulatory approaches that can achieve functional cure of HBV, defined as loss of detectable circulating viral surface antigen (HBsAg).”

Mala Maini, PhD, at UCL’s division of infection & immunity, explained, “Chronic hepatitis B virus infection is a major global health problem and the most common cause of liver cancer in the world. The development of novel therapeutic options is crucial to improve patient care. In this study, we aimed to identify a treatment target to directly inhibit the virus while also boosting the immune cells fighting it.”

Cholesterol is a lipid that is consumed as part of the diet, and which can exert multiple functions within different cells of the body. HBV infects the liver, an organ that is highly enriched in cholesterol and well known for limiting local immune responses. Recent studies have shown that ACAT knockdown or inhibition can directly reduce the growth of several tumors including HBV-related HCC. In their newly reported in vitro study, using human liver disease tissue samples, Maini’s lab at UCL showed that in contrast with currently available therapies, the tested ACAT inhibitors, including the oral drug avasimibe, boosted human HBV-specific CD8+ T cells capable of eliminating the hepatitis B virus. The immune-boosting effect was especially striking in T cells found in the HBV-infected liver and within liver cancer, overcoming the local restraints on immune cell function, and allowing the T cells to target both the virus and cancerous cells. ACAT inhibitors such as avasimibe, taken orally, have been shown to be well tolerated as cholesterol-lowering drugs in humans.

“We find that ACAT inhibition drives metabolic re-modelling, resulting in enhanced expansion and functionality of human CD8+ T cells directed against HBV and HCC, sampled directly from the site of disease,” the team noted. “The increase in functional HBV-specific CD8+ T cells was not merely due to the recovery of pre-existing responses, but also their expansion due to enhanced proliferation.”

The Maini group then collaborated with Jane McKeating, PhD, and her team at the University of Oxford, to show that ACAT inhibitors could also block the HBV life cycle directly. ACAT inhibition demonstrated “a clear antiviral effect,” by reducing both extracellular HBsAg and HBV particle production. The investigators pointed out this dual effect cannot be achieved by the antivirals currently used for the treatment of chronic hepatitis B.

The combined study results indicate that ACAT inhibition could offer a unique combination of both antiviral and immunotherapeutic effects, the scientists suggested. “… these data highlight the potential of ACAT inhibitors to modulate HBV antigen burden beyond existing antiviral agents, an important effect complementary to their capacity to enhance antiviral and antitumor T cell immunity.”

Commenting on the findings, Schmidt said, “The cholesterol-modifying drug is already known to be safe in humans and we hope that our study now informs the development of clinical trials combining cholesterol modulation with other immunotherapies. In summary, our findings offer exciting new possibilities for the treatment of patients with chronic viral infections and cancer.”

The authors concluded, “Here, we show that ACAT inhibition has antiviral activity against HBV, as well as boosting protective anti-HBV and anti-HCC T cells … Thus, ACAT inhibition provides a paradigm of a metabolic checkpoint able to constrain tumors and viruses but rescue exhausted T cells, rendering it an attractive therapeutic target for the functional cure of HBV and HBV-related HCC.”