Development of WHO Guideline-Complying CD4 Diagnostic Chip

Hee Sik Shin and Sungyoung Choi

Hanyang University, Seoul, Korea

Keywords: HIV, Automated Cell Counting, Lego-based Smartphone Microscope, Microfluidics, Negative Selection.

Abstract: Accurate HIV diagnosis using current WHO complying diagnostic chips that measure CD4 protein expression

faces challenges due to donor variability in expression levels and difficulties in isolating target cells with high

purity. To overcome these limitations, quantitative measurement of CD4

and CD8

T cells, along with the

effective removal of red blood cells (RBCs) and granulocytes, is essential. We present a self-powered CD4

diagnostic chip platform designed for high0purity target cell separation using a small volume of whole blood.

The CD4 diagnostic chip employs magnetic separation to remove non-target cells through negative selection,

utilizing ferromagnetic particle-filled inlets and a magnet positioned beneath the inlet. Excessive RBCs and

target cells are further separated by size using a microfluidic lattice. This rapid separation process, facilitated

by a degassed polydimethylsiloxane (PDMS) chip, achieves efficient target cell isolation within 20 minutes.

During separation, the LEGO-integrated smartphone microscope records a real-time video, which is analyzed

using Python-based code. The code distinguishes and removes excess granulocytes based on pixel intensity

and precisely counts pure target cells, enabling analysis and potential diagnosis within 1 minute post-

separation. The cD4 diagnostic chip is a simple, precise, and rapid platform requiring minimal blood volume,

compliant with WHO guidelines for HIV diagnosis.

1 INTRODUCTION

Detection of CD4 expression from whole blood is

critical technology for the clinical diagnosis of human

immunodeficiency virus (HIV). Curren diagnostic

methods that comply with World Health Organization

(WHO) guidelines, such as the ‘VISITECT CD4’

assay, typically involve lysing whole blood and

measuring the total CD4 protein expression (Lechiile

et al., 2022). However, these approaches have

significant limitations.

CD4 protein expression levels on CD4

T cells

exhibit significant interindividual variability. As a

result, individuals with lower CD4 expression can be

misdiagnosed, despite having CD4

T cell counts

within the normal range. Furthermore, CD4

expression is not restricted to T lymphocytes;

monocytes also express CD4, potentially

confounding diagnostic accuracy. Additionally, when

whole blood is processed without lysis and traditional

microfluidic methods are used to isolate target T cells,

sample purity issues may arise, further affecting

diagnostic reliability. Positive selection methods for

CD4

T cell isolation inadvertently include CD4-

expressing monocytes, as previously described,

whereas negative selection methods, while yielding

higher purity, may co-isolate excess granulocytes,

given their higher relative abundance in whole blood.

To address these limitations, it is crucial to rely on

quantitative CD4

T cell counts, rather than CD4

expression levels, for unbiased assessments.

Quantitative measurements for HIV diagnosis require

a CD4

T cell concentration under 200 cells/µL and a

CD4/CD8 T cell ratio below 1. However, existing

studies on quantitative CD4+ T cell assessment do not

consistently comply with WHO guidelines (Yeh et al.,

2017).

To meet WHO standards, diagnostic platforms

must achieve detection at concentrations as low as

200 cells/µL using a minimum whole blood sample

volume of 100 µL and involve a preparation-to-assay

time under 5 minutes.

Here, we developed a CD4 diagnostic chip for a

rapid, precise, and simplified HIV diagnostic platform

suitable for point-of-care testing (POCT). Requiring

only 5 µL of whole blood, the CD4 diagnostic chip

detects concentrations as low as 100 cells/µL. Using

a smartphone microscope integrated with a LEGO-

based setup, real-time video capture and Python-

based software enable quantitative cell counting at the

target cell outlet. The distinct pixel intensity

differences between granulocytes and lymphocytes

Shin, H. S. and Choi, S.

Development of WHO Guideline-Complying CD4 Diagnostic Chip.

DOI: 10.5220/0013254700003911

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2025) - Volume 1, pages 175-179

ISBN: 978-989-758-731-3; ISSN: 2184-4305

175

facilitate visual and software-based cell type

differentiation. Our three-step separation process

yields high-purity CD4

and CD8 T

cells with

quantitative counts in under 25 minutes, meeting high

precision and purity requirements with Python-based

data analysis.

2 METHODS

2.1 Design and Development of the

Lego-Integrated Smartphone

Microscope

Lego bricks were utilized as a structural framework to

stabilize the various components of the setup. A

smartphone (iPhone SE) was mounted in an inverted

position and aligned with the 10X eyepiece lens. A

mirror was used to connect the 20X objective lens to

the eyepiece lens. An LED light source was

positioned at the top of the Lego structure. The CD4

diagnostic chips were securely placed on a

mechanical stage integrated into the framework.

2.2 Fabrication of the CD4 Diagnostic

Chip

The CD4 diagnostic chip was designed following

previously reported methods (Shin et al., 2023). The

chip was fabricated using standard SU-8

photolithography and polydimethylsiloxane (PDMS)

molding techniques. To assemble the chip, it was

positioned on top of a cover glass and affixed. The

assembled chips were then stored in an oven at 65°C

for future use.

2.3 CD4 Diagnostic Chip Procedure

The entire chip was degassed for 20 minutes in a

vacuum chamber, with the outlet reservoirs covered

by cover glass. For long-term storage and portable use,

the chip could be packaged in an aluminum vacuum

seal (yeh et al., 2017). Whole blood was collected in

10 mL BD EDTA Vacutainer tubes, then diluted 1:10

in 0.1% BSA. The blood was mixed with either a

human CD4

T cell isolation kit or a human CD8

cell isolation kit and incubated for 15 minutes at 4°C.

Upon removal of the chip from the chamber, 40 μL of

the blood mixture and 100 μL of 0.1% BSA buffer

were pipetted into the respective blood and buffer

inlets. To assess the purity, precision, and recovery of

the target cells, CD45 (FITC) and either CD4 (APC)

or CD8 (APC) antibodies were added to 100 μL of

0.1% BSA before being loaded onto the chip. After 30

minutes, the target cell outlet was imaged using a

fluorescence microscope (Ti2, Nikon) equipped with

a motorized stage and a scientific CMOS camera

(Andor Inc.).

2.4 Characterization of the Cell

Separation Process

Whole blood samples were serially diluted from 1X

to 1/16X concentrations, and cell separations were

performed using CD4 diagnostic chips. All processes

were recorded in slow-motion mode using a

smartphone, with video frames focused on the region

immediately preceding the target outlet. Post-capture,

cell identification and quantification were conducted

using Python-based code.

3 RESULTS

3.1 Characterization of Target Cell

Separation in the CD4 Diagnostic

Chip

To minimize cost and simplify construction, we

designed an optical setup using a smartphone,

integrated with a modular LEGO-based framework

(Fig. 1A). The CD4 diagnostic chip is positioned on

the objective lens and fixed with a mechanical stage,

enabling cell imaging via smartphone as they flow

through the device. Target cells are isolated by

loading 40µL of whole blood sample and 100µL of

0.1% BSA buffer into the blood and buffer inlets,

respectively.

We further enabled detection of the marker of the

target cells separated from the CD4 diagnostic chip by

imaging the chip’s outlet. This was achieved by adding

a fluorescent antibody to the 0.1% BSA buffer prior to

loading it onto the chip (Fig. 1B). Non-target cells in

the blood sample are magnetically labeled with a

negative selection kit reagent. Upon loading the blood

sample, target cells and non-labeled red blood cells

(RBCs) pass through the microfluidic lattice, while

labeled non-target cells are trapped in the blood inlet,

which contains ferromagnetic particles and a magnet

positioned beneath the blood inlet. The RBCs are

directed to a waste outlet, while target cells are washed

with buffer and sorted into the target outlet (Fig. 1C).

Measurement of CD4

and CD8

T cell separation

throughput over time revealed a consistent decline after

sample loading. This is because of the gradual

reduction in negative pressure over time during the

BIODEVICES 2025 - 18th International Conference on Biomedical Electronics and Devices

176

PDMS degassing process (Fig. 1D).

We then evaluated the effectiveness of target cell

separation from whole blood by assessing the purity

and precision of the sorted cells. Purity was calculated

as the ratio of target cells to the total cell count in the

target cell outlet, while precision was calculated as the

ratio of target cells in the target cell outlet to the total

number of target cells across both outlets. As a result,

we achieved successful separation using CD4

diagnostic chips, obtaining an average purity of

91.7% and precision of 86.2% for CD4

cells, and a

purity of 86.2% and precision of 84.8% for CD8

cells,

all within 30 minutes (Fig. 1E). Additionally, we

assessed the recovery rates of target cells, which were

27.0% for CD4

T cells and 21.7% for CD8

T cells

(Fig. 1F). Recovery was calculated as the ratio of

target cells collected in the target cell outlet to the

total number of target cells loaded at the inlet.

Because of the vacuum pump driven system, the

separation throughput gradually decreased over time.

However, throughput could be enhanced by simply

reducing the volume of the blood sample.

Figure 1: Separation of CD4

and CD8

T cell using the CD4 diagnostic chip. (A) Image of the smartphone microscope setup

with Lego-integrated components. (B) Fluorescence images of sorted CD4

and CD8

T cell. (C) Fluorescence image

depicting microfluidic washing process. (D) Throughput of CD4

and CD8

T cell separation flux over time. (E) Purity and

precision of sorted CD4

and CD8

T cell. (F) Recovery rates of sorted CD4

and CD8

T cell. (H) Repeatability analysis of

the CD4 diagnostic chip for separating CD4

and CD8

T cell.

Development of WHO Guideline-Complying CD4 Diagnostic Chip

177

Figure 2: Quantification of CD4

and CD8

T cell following separation via the CD4 diagnostic chip. (A) Fluorescence imaging

of sorted CD4

T cell and (B) CD8 T

cell, showing residual granulocytes. (C) Plot of pixel intensity differences distinguishing

cell types. (D) Comparison of CD4

and CD8

T cell purity assessed via Python-based automated counting versus manual

counting over a 1-minute video segment. (E) Evaluation of Python code and manual counting methods. (F)Ratio comparison

of CD4

and CD8

T cell between Python-based automated code counting and FACS analysis. (G) Quantitative counts of

CD4

T cells and (H) CD8

T cells obtained from serially diluted whole blood samples, as analyzed by Python code.

We also investigated the repeatability of the CD4

diagnostic chips to assess potential performance

variability. Utilizing whole blood samples from the

same donor, we processed undiluted (1X) and 1/4X

diluted whole blood supplemented with 0.1% BSA

and quantified the CD4

and CD8

cell counts at the

garget cell outlet (Fig. 1G). The results demonstrated

no significant differences in chip performance across

measurements, indicating high repeatability.

3.2 Detection of Target Cell with

Python-Based Code in the CD4

Diagnostic Chip

To detect CD4

T cells at concentrations below 100

cells/µL, it is essential to achieve precise separation,

eliminating non-target cells entirely. However, some

granulocytes persist in the target outlets alongside

CD4

and CD8

T cell (Fig. 2A and 2B).

BIODEVICES 2025 - 18th International Conference on Biomedical Electronics and Devices

178

Given that target cells are isolated using a negative

selection kit, donor variability influences cell

separation outcomes. Although the purity of CD4

and CD8

T cell approaches 90%, the remaining 10%

impurity arises from non-target cells that are not fully

eliminated. In donors with a high abundance of

granulocytes or reduced expression of granulocyte-

specific markers, these granulocytes are inadvertently

collected in the target cell fraction, thereby increasing

the impurity. Notably, brightfield imaging reveals a

significant difference between lymphocytes and

granulocytes. Granulocytes exhibit lower pixel

intensity due to granules, appearing as dark spots, in

contrast to the more transparent appearance of

lymphocytes (Fig. 2C).

The Python-based code, capable of distinguishing

lymphocytes from granulocytes based on pixel

intensity, was utilized to analyze 1-minute segmented

videos recorded via smartphone to assess the purity of

CD4

and CD8

T cell separation. A comparison

between manual counting and Python-based

automated counting methods revealed no significant

differences between the two approached (Fig. 2D). To

further evaluate the precision of cell counting, a

comparative analysis of manual and automated

methods was conducted, demonstrating that the

Python-based code provides accurate and reliable

counts, with no significant discrepancies observed

(Fig. 2E). Given the clinical importance of the CD4

CD8

T cell ratio for diagnosing HIV patients, we

compared the Python-based automated counting

method to conventional FACS analysis. The high

correlation (R

=0.95) between the two methods

indicates strong agreement and confirms the

reliability of the automated approach (Fig. 2F).

To assess the performance of the CD4 diagnostic

chip, whole blood was serially diluted from 1X to

1/16X. Videos were analyzed using the Python-based

code, with recordings segmented into 1-minute

intervals starting from the initiation of the separation

process. We observed a strong correlation between the

concentration of CD4

T cells in PBMCs and the

number of target cells counted using the automated

method, yielding and R

value of 0.93 (Fig. 2G).

Similarly, the concentration of CD8

T cells in

PBMCs showed a high correlation with the automated

cell count, also with an R

value of 0.93 (Fig. 2H).

4 DISCUSSION

We developed a CD4 diagnostic chip integrated with

a LEGO-based smartphone microscope platform,

offering a cost-effective, precise, and straightforward

solution for target T cell separation. The Python-based

code enables accurate counting of target cells at

concentrations below 100 cells/μL, using a minimal

blood volume of just 5 μL. The entire separation and

counting process is completed within 30minutes,

making it well-suited for point-of-care (POC)

applications.

REFERENCES

Lechiile, K., Leeme, T. B., Tenforde, M. W., Bapabi, M.,

Magwenzi, J., Maithamako, O., ... & Jarvis, J. N. (2022).

Laboratory evaluation of the VISITECT advanced

disease semiquantitative point-of-care CD4 test. JAIDS

Journal of Acquired Immune Deficiency

Syndromes, 91(5), 502-507.

Watkins, N. N., Hassan, U., Damhorst, G., Ni, H., Vaid, A.,

Rodriguez, W., & Bashir, R. (2013). Microfluidic CD4+

and CD8+ T lymphocyte counters for point-of-care HIV

diagnostics using whole blood. Science translational

medicine, 5(214), 214ra170-214ra170.

Shin, H. S., Park, J., Lee, S. Y., Yun, H. G., Kim, B., Kim,

J., ... & Choi, S. (2023). Integrative Magneto‐

Microfluidic Separation of Immune Cells Facilitates

Clinical Functional Assays. Small, 19(43), 2302809.

Development of WHO Guideline-Complying CD4 Diagnostic Chip

179