Microsensor Arrays for Saliva
DAVID R. WALT,aTIMOTHY M. BLICHARZ,aRYAN B. HAYMAN,a
DAVID M. RISSIN,aMICHAELA BOWDEN,aWALTER L. SIQUEIRA,b
EVA J. HELMERHORST,bNERLINE GRAND-PIERRE,b
FRANK G. OPPENHEIM,bJASVINDER S. BHATIA,c
FR´ED´ERIC F. LITTLE,cAND JEROME S. BRODYc
aDepartment of Chemistry, Tufts University, Medford, Massachusetts 02155, USA
bDepartment of Periodontology and Oral Biology, Boston University School
of Dental Medicine, Boston, Massachusetts 02118, USA
cDepartment of Medicine, Boston University School of Medicine, Boston,
Massachusetts 02118, USA
ABSTRACT: Optical fiber microarrays have been used to screen saliva
from patients with end-stage renal disease (ESRD) to ascertain the ef-
ficacy of dialysis. We have successfully identified markers in saliva that
ers have been converted to disposable test strips such that patients may
one day be able to monitor their clinical status at home. Details of these
obstructive pulmonary disease (COPD) patients is being screened for
useful diagnostic markers. Our goal is to develop a multiplexed assay for
these protein and nucleic acid biomarkers for diagnosing the cause and
severity of pulmonary exacerbations, enabling more effective treatment
to be administered. These results are reported in the second part of this
KEYWORDS: noninvasive diagnostics; end-stage renal disease; asthma;
COPD; antibody array; DNA array
Saliva has been used increasingly as a sample matrix for systemic disease
diagnosis, based on the premise that saliva reflects the composition of blood
and is a window to an individual’s general health.1,3,4Our interdisciplinary
project team is focused on developing point-of-care diagnostic systems for
common disease states. We initially screen for potentially useful biomarkers
Address for correspondence: David R. Walt, Department of Chemistry, Tufts University, 62 Talbot
Avenue, Medford, MA 02155. Voice: 1-617-627-3470; fax: 1-617-627-3443.
Ann. N.Y. Acad. Sci. 1098: 389–400 (2007). C ?2007 New York Academy of Sciences.
390ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
using standard assays and then transition to microsphere-based assays for any
potentially useful analytes. These microsphere-based sensors and probes are
then integrated into a multiplexed detection platform. This article describes
the implementation of these strategies for two disease case studies to produce
tests that may be used in point-of-care diagnostics.
Salivary Analysis of End-Stage Renal Disease (ESRD)
ESRD is a condition in which kidney functions are severely compromised.
Patients with ESRD require kidney transplantation or frequent hemodialysis
to prevent clinical complications or death due to the buildup of waste pro-
ducts in the blood.2It is critical to monitor kidney function in pre-ESRD
patients to diagnose conversion to the acute disease state. We examined nu-
merous potential renal function biomarkers in the saliva of ESRD patients.
These patients should be an ideal study cohort because the concentration of
some blood analytes decreases dramatically during dialysis. A noninvasive,
self-administered, and rapid method for monitoring kidney function could re-
duce the need for periodic hospital visits and blood testing for pre-ESRD
patients and could potentially be used to evaluate dialysis efficacy for ESRD
Initial ESRD Biomarker Screening Study
ESRD patients in various states of disease progression were enrolled at the
dialysis clinic of Boston University Medical Center (BUMC) and were asked
to donate saliva before and after undergoing dialysis on a weekly basis for
a 2-month period. A panel of candidate analytes was screened for consistent
performed for sodium (Na+), potassium (K+), magnesium (Mg2+), calcium
uric acid (UA), amylase, lactoferrin, esterase, total protein, nucleic acids, and
glucose levels (TABLE 1). Analytes shown in italics exhibited differences be-
tween pre- and post-dialysis saliva composition in initial screening. Analytes
listed in bold type (NO2−, Cl−, Na+, and UA) showed the best correlations
and were monitored in a more extensive study by collecting saliva samples at
regular time intervals throughout dialysis.
ESRD Salivary Biomarker Monitoring During Dialysis
To determine whether these analytes were good indicators for monitoring
the efficacy of dialysis, a study was conducted where saliva samples were
collected from ESRD patients immediately before and after dialysis, as well
WALT et al. 391
TABLE 1. Salivary analytes initially screened for dialysis correlation
Analytes deemed not useful Analytes deemed potentially useful
NOTE: Analytes in italics showed differences between pre- and post-dialysis saliva composition in
initial screening. Analytes in boldface showed the best correlations; for more detail see text.
as at hourly intervals throughout treatment. Saliva levels of NO2−and UA
consistently tracked dialysis, exhibiting decreasing concentrations throughout
the process; the rate of decrease, however, varied by individual. These two
analytes were selected for further evaluation.
Salivary Test Strips for Monitoring Renal Disease
A simple colorimetric test strip was developed to semiquantitatively de-
termine concentrations of NO2−and UA in saliva. This approach offers the
potential for a low-tech and low-cost method for monitoring renal status.
Chromatography paper was impregnated with the NO2−and UA detection
chemistries were followed by adhesion of the test papers onto a vinyl support
material. The colorimetric test paper for salivary NO2−determination is based
Brief immersion in solution produced test pad color intensities proportional
to the concentrations of NO2−and UA in the sample. We developed a cali-
bration color chart to visually determine the concentrations of these analytes
(FIG. 1). FIGURE 2 demonstrates the color change of the test strips after im-
mersion in archived saliva supernatant samples collected from pre- and post-
dialysis ESRD patients.
Following this proof-of-principle study, test strips were employed at the
BUMC Dialysis Clinic for point-of-care salivary NO2−and UA determina-
tions. Stimulated, whole saliva was collected from 19 ESRD patients both be-
fore and after dialysis and was tested immediately using the NO2−/UA strips.
similarly analyzed using the test strips. The outcome confirmed our earlier re-
sults; the test strips could be used to follow NO2−and UA concentrations
during dialysis (FIG. 3).
392ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 1. Photograph of a NO2−/UA test strip used to determine the concentrations
of NO2−(top pad) and UA (bottom pad) in a saliva sample.
Salivary Analysis of Pulmonary Inflammatory Diseases
Pulmonary inflammatory diseases, such as asthma and chronic obstruc-
FIGURE 2. Test strips following immersion in archived saliva supernatant samples
dialysis treatment. Note the comparative difference in test strip color intensity between pre-
and post-dialysis saliva samples.
WALT et al. 393
FIGURE 3. Test strip results compiled by examining stimulated whole saliva samples
from ESRD patients in the BUMC Dialysis Clinic and healthy controls (not dialyzed). Each
strip was evaluated by two analysts and the two concentration readings were averaged.
affects nearly 20 million Americans and costs $11.5 billion in direct expendi-
tures in 2004.9COPD affects nearly 16 million Americans, with another 14
million estimated as living with undiagnosed disease.10
monary inflammatory diseases. The pathogenesis of pulmonary inflammation
irritants, heat/cold/humidity, and bacterial or viral infection.13,14A platform
capable of simultaneously monitoring both the causes of exacerbation as well
as the levels of numerous biomarkers resulting from the pathogenic response
would be a powerful tool for elucidating the differences associated with the
different causes of exacerbation (extrinsic vs. intrinsic). By monitoring many
created. These individual profiles could be regularly monitored to elucidate
the causes of exacerbation and to evaluate the effectiveness of treatment.
Initial Screening of Salivary Cytokines and Chemokines
To determine the endogenous cytokines and chemokines present at de-
tectable concentrations in saliva, we initially screened a small number of
394ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 4. Representative salivary cytokine and chemokine screening results using
the RayBio Cytokine Array V. Each dark spot visible on the array corresponds to an analyte
present in saliva above the detection limit of the kit. Cytokines and chemokines detected
on the array that are associated with pulmonary inflammatory diseases could be examined
with secondary screening studies.
archived saliva supernatant samples from asthmatics and healthy controls us-
ing Human Cytokine Array V Kits from RayBiotech (Norcross, GA, USA).
These commercially available tests are based on a multiplexed enzyme-linked
immunosorbent assay (ELISA) with chemiluminescent detection and they
provide qualitative results identifying the relative levels of 79 cytokines and
chemokines (FIG. 4). Using this method, we were able to identify a number of
salivary cytokines and chemokines that were examined in greater detail using
secondary quantitative assays.
Secondary Screening of Salivary Cytokines and Chemokines
A number of cytokines and chemokines showing elevated levels with pul-
monary inflammation on the RayBio Cytokine Array V were examined with
quantitative ELISA screening studies. Representative preliminary quantitative
salivary screening results of 12 asthmatic patients and 12 healthy controls
are shown in FIGURE 5. When examined closely, no single analyte correlates
WALT et al. 395
FIGURE 5. Representative quantitative screening results for (A) IP-10, (B) VEGF,
(C) IFN-?, and (D) RANTES for 12 asthmatic patients and 12 healthy controls determined
using microtiter plate-based ELISA. The average for each population (asthmatic or healthy
control) is represented by the dark gray bar on the right of each data set.
with pulmonary inflammatory state, but elevated levels of multiple analytes
are present in most of the patients. Cytokines and chemokines showing po-
tential correlations with pulmonary inflammation will be further investigated
to confirm their utility as asthma biomarkers. Finally, by examining numer-
ous inflammatory proteins simultaneously using a multiplexed assay, we hope
to develop a better understanding of the different ways in which asthma can
manifest itself in different patients.
Development of a Multiplexed Fiber Optic Microsphere-Based
Assays for cytokines or chemokines shown to have potential correlation
with pulmonary inflammatory disease or exacerbation could be converted
to microsphere-based probes and pooled for multiplexed screening studies.
To perform multiplexed microsphere-based fiber optic measurements, amine-
oclonal antibodies via glutaraldehyde chemistry. Probes recognizing different
analytes were pooled and deposited into the wells of a fiber optic array to
396 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
produce a multiplexed immunoassay. Additional cytokine and chemokine
probes can be included on the multiplexed array by modifying the compo-
sition of the microsphere bead pool. The current iteration of the multiplexed
Salivary Analysis of Pulmonary Pathogens
We hypothesize that there is a natural exchange of bacteria and viruses
associated with pulmonary exacerbations between upper respiratory tract flu-
ids and saliva. We have identified a variety of organisms and are develop-
ing multiplexed bead-based fiber optic sensor arrays to screen saliva samples
for these pathogens. Probe sequences specific to polymerase chain reaction
(PCR) amplicons for the pathogens listed in TABLE 2 have been incorporated
into Sentrix and BeadChip arrays (Illumina, Inc., San Diego, CA, USA), con-
taining 96 and 16 bundles of 50,000 bead sensors, respectively. Oral control
microorganisms have also been included to verify the validity of our detection
Our present nucleic acid detection approach involves PCR amplification of
the pathogen sequences followed by direct hybridization to oligonucleotide
able extracted bacterial DNA. Whole saliva samples were first centrifuged to
separate cells and particulate matter. Nucleic acids were then isolated and pu-
Qiagen Inc.) and amplified using asymmetric PCR. Hybridization of the PCR
amplicons to the oligonucleotide arrays was detected by staining the biotiny-
lated primers with streptavidin-Cy3. Fluorescence intensities of PCR products
from asthmatic patients versus healthy control samples for Actinomyces naes-
lundii are presented in FIGURE 6.
TABLE 2. Pathogens included in Illumina direct hybridization arrays
BacteriaViruses Oral controls
Respiratory syncytial virus
Parainfluenza 1, 2, 3, 4 a/b
Influenza A virus
Influenza B virus
Rhinovirus A, B
WALT et al. 397
FIGURE 6. FluorescenceintensitiesofActinomycesnaeslundii-001probesfor22asth-
matics (dark gray, left) versus the average of 20 healthy controls (light gray, right) deter-
mined using multiplexed direct hybridization experiments. The limit of detection of this
assay was determined to be 568 a.u.
A threshold was set to the mean of the controls plus one standard devia-
tion. Interestingly, higher intensities were observed for this oral control organ-
ism in asthmatic saliva samples (13/22, 59%) than in control patient samples
tended to show higher fluorescence intensity for asthmatic samples (17/22,
77% above threshold) than for controls (2/20, 10%), as seen in FIGURE 7.
When the patient population was limited to asthmatics with COPD, the results
were similar (data not shown).
has several advantages over current Taqman or immunoassay-based methods.
Hybridizing PCR amplification products to the arrays incorporates an added
level of specificity. The built-in redundancy of ∼30 beads per fiber bundle
ensures that these hybridization events are statistically significant. Conven-
tional culture-based methods are typically slow and unable to differentiate be-
tween strains or serotypes. Direct detection of nucleic acids from pulmonary
pathogens should provide a more accurate profile of current infection than
clinical diagnostic kits that detect antibodies to respiratory pathogens because
of the inherent delay in immune system response and continued antibody pro-
duction following infection clearance.
398 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
FIGURE 7. Haemophilus influenzae-011 probe signals are higher for a majority of the
asthmatics (dark gray, left) relative to controls (light gray, right).
Incorporation of Microfluidics
Sample collection, handling, and pretreatment encompass a wide range of
challenges for a point-of-care device, especially in saliva diagnostics. All in-
terfaces between the device and saliva samples must be cleaned between sam-
pling events to eliminate the possibility of interpatient sample contamination.
We are developing disposable microfluidics cassettes that will incorporate all
necessary extraction, concentration, amplification, and detection chemistries.
Embedded arrays will be included in the cassettes and will contain bead-based
sensors analogous to the BeadChip design used for analyte screening. On
the basis of our experience that hybridization kinetics and limits of detection
are improved by the agitation of samples across sensors,16the microfluidics
platform will also employ thermopneumatic flow oscillation in the detection
In the studies reported here, we have analyzed saliva from ESRD patients
and asthmatics to determine whether this sample matrix could be used for
systemic disease diagnosis and monitoring. For ESRD, we were able to iden-
tify two salivary analytes (NO2−and UA) that were elevated in predialysis
WALT et al. 399
patients and were shown to be reduced following dialysis. Detection
chemistries for these two analytes have been converted to a colorimetric test
strip format for the rapid and facile semiquantitative determination of NO2−
and UA in saliva. The test strips have notable advantages over solution-based
screening methods, namely the ability to provide instantaneous measurements
for two analytes simultaneously in undiluted saliva without expensive instru-
ity of life for ESRD and especially for pre-ESRD patients, as these individuals
could monitor their salivary analyte levels at home, thereby eliminating pe-
riodic visits to the clinic and/or invasive blood testing. For asthma, our goal
is to elucidate the complex network of proteins and pathogens implicated in
pulmonary exacerbations using whole saliva as a diagnostic fluid. By incorpo-
rating assays for promising pulmonary inflammation biomarkers into a mul-
tiplexed point-of-care platform for saliva, physicians would be able to make
better-informed decisions about the cause of exacerbation and appropriate
This work was supported by Grant No. U01 DE14950 from the National
Institute of Dental and Craniofacial Research (NIDCR).
1. MUKHOPADHYAY, R. 2006. Devices to drool for. Anal. Chem. 78: 4255–4259.
2. NATIONAL KIDNEY FOUNDATION. 2006. Chronic Kidney Disease (CKD). http://
www.kidney.org/kidneydisease/ckd/index.cfm. Accessed on November 15,
4. MALAMUD, D. 1992. Saliva as a diagnostic fluid. Br. Med. J. 305: 207–208.
5. FEIN, H., M. BRODERICK, X. ZHANG, et al. 2003. Measurement of nitric oxide
production in biological systems by using Griess reaction assay. Sensors 3: 276–
ASSIGNEE. 1986. U.S. patent 4,631,255. Date of application: December 23.
of lavender Cu(I)-2,2?-bicinchoninate chelate. Clin. Chem. 16: 536.
8. LEE, T.Y., Y.C. LEI, S.-Y. SHEU, et al. INVENTORS; DEVELOPMENT CENTER FOR
BIOTECHNOLOGY, ASSIGNEE. 2004. U.S. patent 6,699,720. Date of application:
9. AMERICAN LUNG ASSOCIATION. July 2006. Trends in asthma morbidity and mor-
10. COPD INTERNATIONAL. COPD information and support. http://www.copd-
international.com/. Accessed on October 19, 2006
400ANNALS OF THE NEW YORK ACADEMY OF SCIENCES
11. KIPS, J.C. 2001. Cytokines in asthma. Eur. Resp. J. 34(Suppl.): 24S–33S.
Proc. Am. Thorac. Soc. 1: 109–114.
14. WEDZICHHA, J. 2004. Role of viruses in exacerbations of chronic obstructive pul-
monary disease. Proc. Am. Thorac. Soc. 1: 115–120.
15. BLICHARZ, T.M. & D.R. WALT. 2006. Detection of inflammatory cytokines using a
fiber optic microsphere immunoassay array. Proc. SPIE- Int. Soc. Optic. Engng.
with an optical imaging microarray capable of attomolar target DNA detection.
Anal. Chem. 77: 5583–5588.