Morphology of Red Blood Cells: A new pathological marker

Morphology of Red Blood Cells: A new pathological marker

Organized by

Debasish Sarkar, Associate Professor

Department of Chemical Engineering

University of Calcutta

Organized byCUChEAA

My initiation into Biological Sciences: intuitive yet rewarding

Fast forward 15 years

Let’s model some cell population!

My earliest interpretation of a cell

And finally, my research started.

Intermediate (and improving)

Concepts further refined

Overview of Erythrocyte/RBC membrane

Protein: 50%; Phospholipid: 20%Cholesterol: 20%; Carbohydrate: 10%

What it looks like

Composition

Cholesterol: 20%; Carbohydrate: 10%

Overview of structural design

Structural details

Morphologically altered erythrocytes

Transformation majorlyinvolves dysfunction ofproteins

� Spectrin

� Ankyrin

� Band 3

� Protein 4.2

Major morphological types

(a) Normal discocyte(b) Echinocyte (with thorny projections)(c) Stomatocyte (with irregular central

stomata)(d) Spherocyte (sphere-shaped)

Mutually exclusive pathways

Importance of morphological study

� RBC morphology acts as a pathological index for several diseases

� Hemoglobinopathies

� Membranopathies

� Infectious diseases

� Inflammatory diseases

� Congenital diseases CANCER

ANAEMIA

ANAEMIA

DIABETES MELLITUS

DIABETES MELLITUSSCD

SCD

PKAN

SNAKE BITE

� Study of RBC morphology is necessary in

� Neonatology

� Cytoskeletal studies

� Membrane composition studies

� Drug interaction studies for research

� Clinical and screening purpose

ANAEMIA

DIABETES MELLITUSSCDPKAN

RENAL DISEASES

SNAKE BITE

SNAKE BITE

Thalassemia

Example of morphological alterations

Dengue (echinocytes) Leukemia (stomatocytes)Patamatamku et al., Asian Biomedicine 11 (2017) 49 - 53 Ohsaka et al., NKGZ 52 (1989) 7-17

7

COVID-19 (echinocytes) Low cholesterol (stomatocytes)

Lakhdari et al., medRxiv 2020Andolfo et al., AJH 91 (2018) 107-121

Measurement of population morphology: Orthodox approach

MICROSCOPIC ANALYSIS

� Laborious

� Subjected to personal bias

� Repeated counting is necessary

Glass (coverslip or slide) induced echinocytosisConfocal microscopy image echinocytosis

Outcome: Qualitative report

Confocal microscopy image

Remedy: Optical measurements based on LASER scattering (Flow cytometry)

FLOW CYTOMETER

• Cells suspended in a fluid• One cell at a time through a laser beam• Tens of thousands of cells rapidly processed• Regular applications in basic research, clinical practice, and clinical trials

ii MSxMI ∑=

Outline: From Flow Cytometry to population morphology

Outcomes of Flow Cytometry Microscopic image

MI: Morphological Index

ii∑

A Morphological scale

Distribution parameters

(mean, median, std. dev., etc.)

Black box/regression

model

Rapid morphological

assay

Morphological scale

Empirical scale and semi-empirical scoring scales

Example of Empirical scale:

Morphological Score (MS)

Based on the classification and scoring system of Besis

et al. (1973)

Mutually exclusive poikilocytosis(supported by bilayer couple

hypothesis)

S. Svetina and B. Zeks, BBA 42 (19823) 586-590M. Bessis and L.S. Lesin, Blood 36 (1970) 399-405

Work on empirical scoring scale

Work on empirical scoring scale

E5M (echinocytic agent)

(stomatocytic agent)

Confocal microscopy and MI estimationOutcomes of flow cytometry

Work on empirical scoring scale

Multivariable regression model(R2 = 0.91 for patient samples)

Yet the scale was purely empirical!

Work on semi-empirical scoring scale

Objective: To introduce some scientific basis in morphological scoring

rms roughnessMeasured throughAFM (tapping mode)AFM schematic

Force as function to tip-sample distance

AFM tip and cantilever assembly

Roughness calculation

10

12 Rrms, MD = 8.11SD = 1.69skewness = 1.13

Work on semi-empirical scoring scale

Roughness measurement of erythrocyte

Our work: Rrms was repeatedly (n = 13) estimated over 1.0 × 1.0 µm2

Girasole et

al. (2007)

Background:

No. of cells0 1 2 3 4 5 6 7 8 9 10 11

Rrm

s, M

D (nm

)

0

2

4

6

8

10

12Rrms, MD = 8.50SD = 0.59skewness = 0.28

Box Index0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Rrm

s (nm

)

0

2

4

6

8estimated over 1.0 × 1.0 µm2

rms roughness of four morphotypes

Designing the semi-empirical scale

Baseline data set of cell and box position averaged rms

roughness

Intermediate scale: X(i)Motive: MSD = 0

Intermediate scale: Y(i)Motive: to induce symmetry

Final MS scaleThrough normalization of Y(i)

Sem

i-em

piric

al M

S s

cale

0.0

0.5

1.0

Sp (% shift: 60)

E (% shift:-20)

SpE (no shift)

Parity between semi-empirical and empirical scales

Empirical MS scale -1.0 -0.5 0.0 0.5 1.0

Sem

i-em

piric

al M

S s

cale

-1.0

-0.5

Trendline

450 lineSpSt (no shift) “………there will always be rich and poor. Rich in gifts, poor in gifts. Rich in love, poor in love…….”: Danilov (Enemy at the Gates)

The model: MI = f (flowcytometric parameters)

ANN model

Model data points

Model exclusivedata points

Model application: CA patients

Confocal Imaging Flowcytometry (width distribution) Parity plot

Results of correlation study: Hematological parameters and MI

CBC parameters:1. TLC 2. HCT3. MCV 4. PLT

Best Intra-group correlation Correlation with MI

Leftovers of empiricity

The problem lies in the scanning protocol

Study by Girasole et al. (2007) Our report (2016)

Choice of fixed scan size was purely empirical; Rrms increases with scan area

• Variation reported by Girasole et al (2007) • Assumption of Area = 1.0 ×1.0 µm2

Hint about “Self Organization” of Rrms

Self-Organization

Self-organization, also called (in the social

sciences) spontaneous order, is a process where some form

of overall order arises from local interactions between parts

of an initially disordered system.

Source: Wikipedia

Framework to describe self-organization

Scaling parameters of Rrms vs. scan size: baseline data for morphological score.

Power law and self organization

Power law in distribution:P(x) ~ x-n

(n: fractal dimension)

Power law with positive exponent: y(x) ~ xHf

and Fd = 3-Hf

The scaling relation

Trimming down the empirical basis

Increasing size of scan elements: 0.25 to 20 µm2

Distribution of Rrms for 0. 5 × 0. 5 µm2

Cell-to-cell

variations of Rrms, MD

Trimming down the empirical basis

Self similarity in our work

Perfectly self similar

• Confidence Interval of Hf for four morphotypes > 95%

• Fd scale: Sp > E > D > St

MS scales and parity

But our Danilov is dead……

Findings

Model: ANN Model formulation data: MI from confocal imaging and flowcytometric data. Outcomes: MI = f (flowcytometric parameters)

The best modelThe least accurate

Findings

Whole data set (n = 50) Model formulation-exclusive

data for CA patients (n = 30)

• Other possible combinations of Xiwere also tested.

• Only X3 = Mean SSC, others unchanged produced R2 = 0.98

A stroke of serendipity….

• Nearly identical profiles for all three abnormal morphotypes

• Patient sample profile (for discocytes) deviates towards the lower side

• Prediction of early signature of the disease (?)

Some of our other works related to hematology….

Soft Matter 14 (2018) 7335-7346.