Developmental plasticity of red blood cell homeostasis

Transformer-based artificial intelligence on single-cell clinical data for homeostatic mechanism inference and rational biomarker discovery

Artificial intelligence (AI) applied to single-cell data has the potential to transform our understanding of biological systems by revealing patterns and mechanisms that simpler traditional methods miss. Here, we develop a general-purpose, interpretable AI pipeline consisting of two deep learning models: the Multi- Input Set Transformer++ (MIST) model for prediction and the single-cell FastShap model for interpretability. We apply this pipeline to a large set of routine clinical data containing single-cell measurements of circulating red blood cells (RBC), white blood cells (WBC), and platelets (PLT) to study population fluxes and homeostatic hematological mechanisms. We find that MIST can use these single-cell measurements to explain 70-82% of the variation in blood cell population sizes among patients (RBC count, PLT count, WBC count), compared to 5-20% explained with current approaches. MIST's accuracy implies that substantial information on cellular production and clearance is present in the single-cell measurements. MIST identified substantial crosstalk among RBC, WBC, and PLT populations, suggesting co-regulatory relationships that we validated and investigated using interpretability maps generated by single-cell FastShap. The maps identify granular single-cell subgroups most important for each population's size, enabling generation of evidence-based hypotheses for co-regulatory mechanisms. The interpretability maps also enable rational discovery of a single-WBC biomarker, "Down Shift", that complements an existing marker of inflammation and strengthens diagnostic associations with diseases including sepsis, heart disease, and diabetes. This study illustrates how single-cell data can be leveraged for mechanistic inference with potential clinical relevance and how this AI pipeline can be applied to power scientific discovery.

Haematological setpoints are a stable and patient-specific deep phenotype

The complete blood count (CBC) is an important screening tool for healthy adults and a common test at periodic exams. However, results are usually interpreted relative to one-size-fits-all reference intervals1,2, undermining the precision medicine goal to tailor care for patients on the basis of their unique characteristics3,4. Here we study thousands of diverse patients at an academic medical centre and show that routine CBC indices fluctuate around stable values or setpoints5, and setpoints are patient-specific, with the typical healthy adult's nine CBC setpoints distinguishable as a group from those of 98% of other healthy adults, and setpoint differences persist for at least 20 years. Haematological setpoints reflect a deep physiologic phenotype enabling investigation of acquired and genetic determinants of haematological regulation and its variation among healthy adults. Setpoints in apparently healthy adults were associated with significant variation in clinical risk: absolute risk of some common diseases and morbidities varied by more than 2% (heart attack and stroke, diabetes, kidney disease, osteoporosis), and absolute risk of all-cause 10 year mortality varied by more than 5%. Setpoints also define patient-specific reference intervals and personalize the interpretation of subsequent test results. In retrospective analysis, setpoints improved sensitivity and specificity for evaluation of some common conditions including diabetes, kidney disease, thyroid dysfunction, iron deficiency and myeloproliferative neoplasms. This study shows CBC setpoints are sufficiently stable and patient-specific to help realize the promise of precision medicine for healthy adults.

The quantum cell

There is a consensus that we are conscious of something greater than ourselves, as if we are derived from some other primordial set of principles. Classical or Newtonian physics is based on the Laws of Nature. Conversely, in a recent series of articles, it has been hypothesized that the cell was formed from lipid molecules submerged in the primordial ocean that covered the earth 100 million years after it formed. Since lipids are amphiphiles, with both a positively- and negatively-charged pole, the negatively-charged pole is miscible in water. Under the influence of earth's gravity, the lipid molecules stand up perpendicularly to the surface of the water, packing together until the negative charge neutralizes the Van der Waals force for surface tension, causing the lipid molecules to 'leap' into the micellar form as a sphere with a semi-permeable membrane. Particles in the water freely enter and exit such spheres based on mass action. Over time such protocells evolved Symbiogenesis, encountering factors that posed existential threats, assimilating them to form physiology to maintain homeostatic control. Importantly, when differentiated lung or bone cells are exposed to zero gravity, they lose their phenotypic identity in their evolved state, which has been interpreted as transiting from local to non-local consciousness.

Rapid and reversible optical switching of cell membrane area by an amphiphilic azobenzene

Cellular membrane area is a key parameter for any living cell that is tightly regulated to avoid membrane damage. Changes in area-to-volume ratio are known to be critical for cell shape, but are mostly investigated by changing the cell volume via osmotic shocks. In turn, many important questions relating to cellular shape, membrane tension homeostasis and local membrane area cannot be easily addressed because experimental tools for controlled modulation of cell membrane area are lacking. Here we show that photoswitching an amphiphilic azobenzene can trigger its intercalation into the plasma membrane of various mammalian cells ranging from erythrocytes to myoblasts and cancer cells. The photoisomerization leads to a rapid (250-500 ms) and highly reversible membrane area change (ca 2 % for erythrocytes) that triggers a dramatic shape modulation of living cells.

Valls R, Pallares Lopez R, Palanques-Tost E, Mow C, Westover MB, Aguirre AD, Higgins JM. Association of Red Blood Cell Distribution Width With Mortality Risk in Hospitalized Adults With SARS-CoV-2 Infection

Importance

Coronavirus disease 2019 (COVID-19) is an acute respiratory illness with a high rate of hospitalization and mortality. Biomarkers are urgently needed for patient risk stratification. Red blood cell distribution width (RDW), a component of complete blood counts that reflects cellular volume variation, has been shown to be associated with elevated risk for morbidity and mortality in a wide range of diseases.

Objective

To investigate whether an association between mortality risk and elevated RDW at hospital admission and during hospitalization exists in patients with COVID-19.

Design, Setting, and Participants

This cohort study included adults diagnosed with SARS-CoV-2 infection and admitted to 1 of 4 hospitals in the Boston, Massachusetts area (Massachusetts General Hospital, Brigham and Women's Hospital, North Shore Medical Center, and Newton-Wellesley Hospital) between March 4, 2020, and April 28, 2020.

Main Outcomes and Measures

The main outcome was patient survival during hospitalization. Measures included RDW at admission and during hospitalization, with an elevated RDW defined as greater than 14.5%. Relative risk (RR) of mortality was estimated by dividing the mortality of those with an elevated RDW by the mortality of those without an elevated RDW. Mortality hazard ratios (HRs) and 95% CIs were estimated using a Cox proportional hazards model.

Results

A total of 1641 patients were included in the study (mean [SD] age, 62[18] years; 886 men [54%]; 740 White individuals [45%] and 497 Hispanic individuals [30%]; 276 nonsurvivors [17%]). Elevated RDW (>14.5%) was associated with an increased mortality risk in patients of all ages. The RR for the entire cohort was 2.73, with a mortality rate of 11% in patients with normal RDW (1173) and 31% in those with an elevated RDW (468). The RR in patients younger than 50 years was 5.25 (normal RDW, 1% [n = 341]; elevated RDW, 8% [n = 65]); 2.90 in the 50- to 59-year age group (normal RDW, 8% [n = 256]; elevated RDW, 24% [n = 63]); 3.96 in the 60- to 69-year age group (normal RDW, 8% [n = 226]; elevated RDW, 30% [104]); 1.45 in the 70- to 79-year age group (normal RDW, 23% [n = 182]; elevated RDW, 33% [n = 113]); and 1.59 in those ≥80 years (normal RDW, 29% [n = 168]; elevated RDW, 46% [n = 123]). RDW was associated with mortality risk in Cox proportional hazards models adjusted for age, D-dimer (dimerized plasmin fragment D) level, absolute lymphocyte count, and common comorbidities such as diabetes and hypertension (hazard ratio of 1.09 per 0.5% RDW increase and 2.01 for an RDW >14.5% vs ≤14.5%; P < .001). Patients whose RDW increased during hospitalization had higher mortality compared with those whose RDW did not change; for those with normal RDW, mortality increased from 6% to 24%, and for those with an elevated RDW at admission, mortality increased from 22% to 40%.

Conclusions and Relevance

Elevated RDW at the time of hospital admission and an increase in RDW during hospitalization were associated with increased mortality risk for patients with COVID-19 who received treatment at 4 hospitals in a large academic medical center network.

Unraveling Disease Pathophysiology with Mathematical Modeling

Modeling has enabled fundamental advances in our understanding of the mechanisms of health and disease for centuries, since at least the time of William Harvey almost 500 years ago. Recent technological advances in molecular methods, computation, and imaging generate optimism that mathematical modeling will enable the biomedical research community to accelerate its efforts in unraveling the molecular, cellular, tissue-, and organ-level processes that maintain health, predispose to disease, and determine response to treatment. In this review, we discuss some of the roles of mathematical modeling in the study of human physiology and pathophysiology and some challenges and opportunities in general and in two specific areas: in vivo modeling of pulmonary function and in vitro modeling of blood cell populations.

A cross-sectional survey based on blood VOCs, hematological parameters and urine indicators in a population in Jilin, Northeast China

The objective of this study was to examine whether long-term exposure to low-dose volatile organic compounds (VOCs) will have an effect on the health of non-occupational population. A total of 499 non-occupational participants aged more than 18 that live around Jilin Petrochemical Industrial Zone were chosen by stratified cluster random sampling. Their blood VOCs' levels, hematological parameters and urine indicators together with detailed questionnaire data were used to find possible relationships using binary logistic regression analysis. The detection rate of benzene in the blood was high in the non-occupational population around the industrial area, and it even reached 82.3% in males but no significant difference was recorded between male and female population. In addition, trichloroethane (male: 33.2% V female: 21.7%; p = 0.002), carbon tetrachloride (males: 20.3% V females: 7.5%; p < 0.001) and trichlorethylene (male: 34.9% V female: 24.7%; p = 0.004) all showed significant differences in gender, and without exception, the prevalence of males was higher in these three VOCs than of females. The changes in red blood cell (RBC), hematocrit (HCT) and basophils are correlated with carbon tetrachloride, trichloroethylene and chloroform, respectively. And RBC, HCT and basophils are statistically significant in male compared with female of the study population. The increase in trichlorethylene was associated with an increase of 1.723% (95% CI 1.058-2.806) in HCT. The increase in carbon tetrachloride showed a more significant correlation with an increase of 2.638% in RBC count (95% CI 1.169-5.953). And trichloromethane led to a 1.922% (95% CI 1.051-3.513) increase in basophils. The changes in urinary WBC, urine ketone (KET) and urinary bilirubin (BIL) showed significant correlation with benzene, carbon tetrachloride and dibromochloromethane, respectively. The correlation in females is more significant than in males. The increase of benzene in the female population increased urinary leukocyte count by 2.902% (95% CI 1.275-6.601). The effect of carbon tetrachloride on KET was particularly pronounced, resulting in an increase of 7.000% (95% CI 1.608-30.465). Simultaneously, an increase in dibromochloromethane caused an increase of 4.256% (95% CI 1.373-13.192) in BIL. The changes in RBC, HCT and basophils can only serve as an auxiliary indicator for disease diagnosis, so they have no significant clinical significance. However, the alteration of urinary WBC, KET and BIL has great clinical significances, and it is suggested that the monitoring of the above indicators from low-dose long-term exposure be strengthen in this area.

The Benefits and Risks of Iron Supplementation in Pregnancy and Childhood

Iron deficiency is the most common micronutrient deficiency in the world and disproportionately affects pregnant women and young children. Iron deficiency has negative effects on pregnancy outcomes in women and on immune function and neurodevelopment in children. Iron supplementation programs have been successful in reducing this health burden. However, iron supplementation of iron-sufficient individuals is likely not necessary and may carry health risks for iron-sufficient and potentially some iron-deficient populations. This review considers the physiology of iron as a nutrient and how this physiology informs decision-making about weighing the benefits and risks of iron supplementation in iron-deficient, iron-sufficient, and iron-overloaded pregnant women and children.

Using Hematology Data from Malaria Vaccine Research Trials in Humans and Rhesus Macaques (Macaca mulatta) To Guide Volume Limits for Blood Withdrawal

Guidelines on safe volume limits for blood collection from research participants in both humans and laboratory animals vary widely between institutions. The main adverse event that may be encountered in large blood volume withdrawal is iron-deficiency anemia. Monitoring various parameters in a standard blood panel may help to prevent this outcome. To this end, we analyzed the Hgb and MCV values from 43 humans and 46 macaques in malaria vaccine research trials. Although the percentage of blood volume removed was greater for macaques than humans, macaques demonstrated an overall increase of MCV over time, indicating the ability to respond appropriately to frequent volume withdrawals. In contrast, humans showed a consistent declining trend in MCV. These declines in human MCV and Hgb were significant from the beginning to end of the study despite withdrawals that were smaller than recommended volume limits. Limiting the volume withdrawn to no more than 12.5% seemed to be sufficient for macaques, and at 14% or more individual animals tended to fail to respond appropriately to large-volume blood loss, as demonstrated by a decrease in MCV. The overall positive erythropoietic response seen in macaques was likely due to the controlled, iron-fortified diet they received. The lack of erythropoietic response in the human subjects may warrant iron supplementation or reconsideration of current blood volume withdrawal guidelines.

Mechanistic modeling of hemoglobin glycation and red blood cell kinetics enables personalized diabetes monitoring

The amount of glycated hemoglobin (HbA1c) in diabetic patients' blood provides the best estimate of the average blood glucose concentration over the preceding 2 to 3 months. It is therefore essential for disease management and is the best predictor of disease complications. Nevertheless, substantial unexplained glucose-independent variation in HbA1c makes its reflection of average glucose inaccurate and limits the precision of medical care for diabetics. The true average glucose concentration of a nondiabetic and a poorly controlled diabetic may differ by less than 15 mg/dl, but patients with identical HbA1c values may have true average glucose concentrations that differ by more than 60 mg/dl. We combined a mechanistic mathematical model of hemoglobin glycation and red blood cell kinetics with large sets of within-patient glucose measurements to derive patient-specific estimates of nonglycemic determinants of HbA1c, including mean red blood cell age. We found that between-patient variation in derived mean red blood cell age explains all glucose-independent variation in HbA1c. We then used our model to personalize prospective estimates of average glucose and reduced errors by more than 50% in four independent groups of greater than 200 patients. The current standard of care provided average glucose estimates with errors >15 mg/dl for one in three patients. Our patient-specific method reduced this error rate to 1 in 10. Our personalized approach should improve medical care for diabetes using existing clinical measurements.

Low-Dose Iron Supplementation in Infancy Modestly Increases Infant Iron Status at 9 Mo without Decreasing Growth or Increasing Illness in a Randomized Clinical Trial in Rural China

BACKGROUND

Previous trials of iron supplementation in infancy did not consider maternal iron supplementation.

OBJECTIVE

This study assessed effects of iron supplementation in infancy and/or pregnancy on infant iron status, illnesses, and growth at 9 mo.

METHODS

Enrollment occurred from December 2009 to June 2012 in Hebei, China. Infants born to women in a pregnancy iron supplementation trial were randomly assigned 1:1 to iron [∼1 mg Fe/(kg · d) as oral iron proteinsuccynilate] or placebo from 6 wk to 9 mo, excluding infants with cord ferritin <35 μg/L. Study groups were pregnancy placebo/infancy placebo (placebo/placebo), pregnancy placebo/infancy iron (placebo/iron), pregnancy iron/infancy placebo (iron/placebo), and pregnancy iron/infancy iron (iron/iron). The primary outcome was 9-mo iron status: iron deficiency (ID) by cutoff (≥2 abnormal iron measures) or body iron <0 mg/kg and ID + anemia (hemoglobin <110 g/L). Secondary outcomes were doctor visits or hospitalizations and weight or length gain from birth to 9 mo. Statistical analysis by intention to treat and dose-response (between number of iron bottles received and outcome) used logistic regression with concomitant RRs and general linear models, with covariate control as applicable.

RESULTS

Of 1482 infants randomly allocated, 1276 had 9-mo data (n = 312-327/group). Iron supplementation in infancy, but not pregnancy, reduced ID risk: RRs (95% CIs) were 0.89 (0.79, 0.998) for placebo/iron compared to placebo/placebo, 0.79 (0.63, 0.98) for placebo/iron compared to iron/placebo, 0.87 (0.77, 0.98) for iron/iron compared to placebo/placebo, and 0.86 (0.77, 0.97) for iron/iron compared to iron/placebo. However, >60% of infants still had ID at 9 mo. Receiving more bottles of iron in infancy was associated with better infant iron status at 9 mo but only among iron-supplemented infants whose mothers were also iron supplemented (i.e., the iron/iron group). There were no group differences in hospitalizations or illnesses and no adverse effects on growth overall or among infants who were iron sufficient at birth.

CONCLUSIONS

Iron supplementation in Chinese infants reduced ID at 9 mo without adverse effects on growth or illness. Effects of iron supplementation in pregnancy were observed only when higher amounts of iron were distributed in infancy. This trial was registered at clinicaltrials.gov as NCT00613717.

Modulation of red blood cell population dynamics is a fundamental homeostatic response to disease

Increased red blood cell (RBC) volume variation (RDW) has recently been shown to predict a wide range of mortality and morbidity: death due to cardiovascular disease, cancer, infection, renal disease, and more; complications in heart failure and coronary artery disease, advanced stage and worse prognosis in many cancers, poor outcomes in autoimmune disease, and many more. The mechanisms by which all of these diseases lead to increased RDW are unknown. Here we use a semi-mechanistic mathematical model of in vivo RBC population dynamics to dissect the factors controlling RDW and show that elevated RDW results largely from a slight reduction in the in vivo rate of RBC turnover. RBCs become smaller as they age, and a slight reduction in the rate of RBC turnover allows smaller cells to continue circulating, expanding the low-volume tail of the RBC population's volume distribution, and thereby increasing RDW. Our results show that mildly extended RBC lifespan is a previously unrecognized homeostatic adaptation common to a very wide range of pathologic states, likely compensating for subtle reductions in erythropoietic output. A mathematical model-based estimate of the clearance rate may provide a novel early-warning biomarker for a wide range of morbidity and mortality.

Fetal iron deficiency and genotype influence emotionality in infant rhesus monkeys

BACKGROUND

Anemia during the third trimester of fetal development affects one-third of the pregnancies in the United States and has been associated with postnatal behavioral outcomes. This study examines how fetal iron deficiency (ID) interacts with the fetal monoamine oxidase A (MAOA) genotype. MAOA metabolizes monoamine neurotransmitters. MAOA polymorphisms in humans affect temperament and modify the influence of early adverse environments on later behavior.

OBJECTIVE

The aim of the study was to advance translation of developmental ID research in animal models by taking into account genetic factors that influence outcomes in human populations.

METHODS

Male infant rhesus monkeys 3-4 mo old born to mothers fed an ID (10 ppm iron) diet were compared with controls (100 ppm iron). Infant monkeys with high- or low-transcription rate MAOA polymorphisms were equally distributed between diet groups. Behavioral responses to a series of structured experiences were recorded during a 25-h separation of the infants from their mothers.

RESULTS

Infant monkeys with low-transcription MAOA polymorphisms more clearly demonstrated the following ID effects suggested in earlier studies: a 4% smaller head circumference, a 39% lower cortisol response to social separation, a 129% longer engagement with novel visual stimuli, and 33% lesser withdrawal in response to a human intruder. The high MAOA genotype ID monkeys demonstrated other ID effects: less withdrawal and emotionality after social separation and lower "fearful" ratings.

CONCLUSION

MAOA × ID interactions support the role of monoamine neurotransmitters in prenatal ID effects in rhesus monkeys and the potential involvement of common human polymorphisms in determining the pattern of neurobehavioral effects produced by inadequate prenatal nutrition.

Red blood cell population dynamics

Hematology analyzers provide a static snapshot of the circulating population of red blood cells (RBCs). The RBC population is rapidly changing, with more than 2 million RBCs turning over every second in the typical healthy adult. The static snapshot provided by the complete blood count does not capture many of the dynamic aspects of this population, such as the rate of RBC maturation and the rate of RBC turnover. By integrating basic science with hematology analyzer measurements, it is possible to estimate the rates of these dynamic processes, yielding new insights into human physiology, with potential diagnostic application.

In vivo volume and hemoglobin dynamics of human red blood cells

Human red blood cells (RBCs) lose ∼30% of their volume and ∼20% of their hemoglobin (Hb) content during their ∼100-day lifespan in the bloodstream. These observations are well-documented, but the mechanisms for these volume and hemoglobin loss events are not clear. RBCs shed hemoglobin-containing vesicles during their life in the circulation, and this process is thought to dominate the changes in the RBC physical characteristics occurring during maturation. We combine theory with single-cell measurements to investigate the impact of vesiculation on the reduction in volume, Hb mass, and membrane. We show that vesicle shedding alone is sufficient to explain membrane losses but not volume or Hb losses. We use dry mass measurements of human RBCs to validate the models and to propose that additional unknown mechanisms control volume and Hb reduction and are responsible for ∼90% of the observed reduction. RBC population characteristics are used in the clinic to monitor and diagnose a wide range of conditions including malnutrition, inflammation, and cancer. Quantitative characterization of cellular maturation processes may help in the early detection of clinical conditions where maturation patterns are altered.