Real-time diagnostics for Pseudomonas: Towards earlier and better

detection

Luke Hoffman
Director, Center for CF Microbiology
University of Washington Pediatrics and Microbiology
June, 2019
 Disclosures

• I participate in investigator-initiated studies supported by Mylan, Novartis

and Vertex

• I direct a US national resource center for the Cystic Fibrosis Foundation

that provides consultation to private and public researchers
 Early, sensitive, accurate identification of P.
aeruginosa infection is an important goal

• Fast, accurate detection could enable:

• Timely initiation of eradication therapy

• Early infection

• Timely initiation of suppressive therapy

• Chronic infection

How do we currently identify P. aeruginosa in our patients?

 Current culture-based P. aeruginosa detection
methods

• Culture on selective media

~48 hours

• Subculture individual colonies

Expectorated,
lavaged, or ~24-48 hours
swabbed
secretions • Species identification (at 3-4 days)
Not even considering susceptibility
testing (at least another day)
 Sampling issues in P. aeruginosa detection

Ease, timing, and accuracy of detection are impacted by

sample type.

• Children rarely expectorate

Expectorated, • Bronchoalveolar lavages are not simple

lavaged, or • Oropharyngeal swabs are not as accurate as we’d like:
swabbed
secretions • Sensitivity 23%, specificity 91% versus BAL
• (Breuer et al., ERJ 2018 and others)
 Other issues in P. aeruginosa detection by
culture

• Culture detection requires at least 103 viable cells

per mL of sample

• Slowly-growing isolates often missed

• Common with chronic CF infection- small-
Expectorated, colony variants
lavaged, or
• Viable but nonculturable cells are missed
swabbed • Common with antibiotic treatment
secretions
• Current methods only sample a small fraction of
the whole, infecting P. aeruginosa population
 Current culture-based P. aeruginosa detection
methods: Recent improvements already in use

• Culture on selective media

~48 hours

• Subculture for individual colonies

Expectorated,
lavaged, or ~24-48 hours
swabbed
secretions • Species identification

Traditionally: Biochemical testing.

 Current culture-based P. aeruginosa detection
methods: Recent improvements already in use
*Matrix-assisted laser desorption/ionization

• Culture on selective media

~48 hours

• Subculture for individual colonies

Expectorated,
lavaged, or ~24 hours
swabbed
secretions • Species identification

MALDI* (rapid proteomic fingerprinting) saves at least one day.

 Current culture-based P. aeruginosa detection
methods: Recent improvements already in use

• Culture on selective media

~48 hours

• Subculture for individual colonies

Expectorated,
lavaged, or ~24 hours
swabbed
secretions • Species identification

PCR is also sometimes used.

 What do we want in a diagnostic test for P. aeruginosa?
• Turnaround time faster than 3 days

• High sensitivity
• Detects P. aeruginosa when present
• Has a limit of detection at least as good as culture

• High specificity
• Has few or no false positives

• Noninvasive

• Inexpensive (MALDI ~$8 USD per sample; PCR ~$150-$200)

• Possibility of susceptibility information would be a plus

 Some promising options for P. aeruginosa
detection
• Serology

• DNA-based:

• Species-specific PCR

• Sequencing

• Metabolites
 Some promising options for P. aeruginosa
detection
• Serology

• DNA-based:

• Species-specific PCR

• Sequencing

• Metabolites
 Serological identification of P. aeruginosa
infection in CF
• Early studies were promising…
 Serological identification of P. aeruginosa
infection in CF
… But subsequent studies have been discouraging.
 Some promising options for P. aeruginosa
detection
• Serology

• DNA-based:

• Species-specific PCR

• Sequencing

• Metabolites
 PCR-based detection of P. aeruginosa

Expectorated, PCR
lavaged, or Extract DNA from isolates
swabbed
secretions

Identify using species-specific genes

 PCR-based detection of P. aeruginosa
Currently done.

Saves… some time…

Expectorated, PCR
lavaged, or Extract DNA from isolates
swabbed
secretions

Identify using species-specific genes

 PCR-based detection of P. aeruginosa

Extract DNA directly from sample

Expectorated, PCR
lavaged, or
swabbed
secretions

Identify using species-specific genes

 PCR-based detection of P. aeruginosa

• Many studies have investigated this approach

• Diverse methods, gene targets, results

• Two excellent reviews:

• Deschaght et al., JCF, 2011 (PCR)

• Pattison et al., JCF, 2013 (all DNA-based)

• One excellent recent study…

PCR-based detection of P. aeruginosa
 PCR-based detection of P. aeruginosa
Multicentre, 3-year study of 96 patients without chronic P. aeruginosa

Compared culture and two-step quantitative PCR (qPCR) in sputum

“free”= >1 year culture-negative

“never”= no prior positive culture

Sensitivity: 94% Compared

Specificity: 86% with culture
Detected median of 8 mos earlier
 Advantages and disadvantages of PCR-based
detection of P. aeruginosa
Advantages:
• Much more rapid turnaround time possible
• But most laboratories must ”batch” to be feasible, limiting this

• Detects viable, nonculturable bacteria

• May be more sensitive than culture

Disadvantages:
• No susceptibility information

• Not validated or feasible for directing CF care

• Prognostic utility not known
• Not feasible for most labs currently
 Advantages and disadvantages of PCR-based
detection of P. aeruginosa
Advantages:
• Much more rapid turnaround time possible
• But most laboratories must ”batch” to be feasible, limiting this

• Detects viable, nonculturable bacteria

• May be more sensitive than culture

Disadvantages:
• No susceptibility information

• Not validated or feasible for directing CF care

• Prognostic utility not known
• Not feasible for most labs currently
Current PCR-based bacterial detection methods
Rapid, approved multiplex PCR methods available

Pneumonia panels currently include P. aeruginosa

No panel specifically for CF

Not yet considered commercially viable for CF sputum

 Advantages and disadvantages of PCR-based
detection of P. aeruginosa
Advantages:
• Much more rapid turnaround time possible
• Most laboratories must ”batch” to be feasible

• Detects viable, nonculturable bacteria

• May be more sensitive than culture

Disadvantages:
• No susceptibility information

• Not validated or feasible for directing CF care

• Prognostic utility not known
• Not feasible for most labs currently
 Prognostic utility of P. aeruginosa qPCR:
A recent example

P. aeruginosa detected by PCR in sputum in the year prior to positive culture

- Tested whether higher P. aeruginosa DNA load had prognostic significance

The quantitative results provided did not predict:

- Detection in subsequent culture
- Likelihood of eradication by inhaled tobramycin

Therefore: Earlier detection, but with unclear clinical benefit

 Sequencing-based detection of P. aeruginosa

Extract DNA directly from sample

Expectorated,
lavaged, or
swabbed
secretions
Sequencing
(I’ve skipped a few critical steps)
 Sequencing-based detection of P. aeruginosa:
Two excellent recent studies

• Performed sequencing on archived, previously cultured CF sputum

• Correlation with concurrent culture particularly good for P. aeruginosa
• Also detected P. aeruginosa in culture-negative samples
 Sequencing-based detection of P. aeruginosa:
Advantages and disadvantages
Advantages:
• Simultaneous analysis of all detectable bacteria

• Quantitative data (at least relative)

• The possibility of antibiotic-resistance information

• The possibility of strain-level information

Disadvantages:
• Equipment and expertise is not routinely available to any lab

• Like PCR, not validated against clinical outcomes

 Sequencing-based detection of P. aeruginosa:
Advantages and disadvantages
Advantages:
• Simultaneous analysis of all detectable bacteria

• Quantitative data (at least relative)

• The possibility of antibiotic-resistance information

• The possibility of strain-level information

Disadvantages:
• Equipment and expertise is not routinely available to any lab

• Like PCR, not validated against clinical outcomes

 Sequencing-based detection of P. aeruginosa:
Two excellent recent studies

• Proof of principle that rapid, multispecies detection by molecular methods

can be feasible in a busy clinical laboratory.
 An issue common to all DNA-based methods

DNA is present in live and dead cells

Extracellular DNA is also present

P. aeruginosa DNA seems to be particularly

troublesome
An issue common to all DNA-based methods
 An issue common to all DNA-based methods
Nelson, Cell Reports, 2019

Digesting freeFigure
DNA in CF sputum
3. Sequence-Base Figure
d Phylogenetic
DNA extracted from each ofDNA
4 DTT-treated,
decreases
3. Sequence-Base
extracted from
Composition
homogenized
d Phylogenetic
each of 4 DTT-treated,
estimated
of DTT-Treated
Composition
CF sputum samples
homogenized
was analyzed
CF sputum
P.samples
using
aeruginosa
Test Setof2 DTT-Treated
Samples afterTest
Standard
Set 2 Samples
both metagenomic
abundance
and Benzonase
was analyzedsequencing
after Standard
Extraction
using both metagenomic
and Benzonase
(MetaPhlAn2, top
sequencing
row) and(MetaP
16S amplicon sequencing (16S
16S Amplicon,
amplicon sequencing
bottom row).
(16S Amplicon, bottom row).
See also Tables S1–S4. See also Tables S1–S4.

P. aeruginosa DNA is common in CF sputum

smaller set of 4 sputumsmaller
samples set(‘‘test set 2’’), samples
of 4 sputum (‘‘test
1 from test 1 2’’),
set set tion1 methods
from test when
set 1 applied
tion methods
to test set
when
1, suggesting
applied to test
that set
this1,ef-sugge
(186), and 1 sample each (186),
fromand31additional
sample eachCF patients.
from 3 additional
Each fect CF was
patients.
minimal
Each(Tablefect
S1).
was16Sminimal
amplicon
(Table
sequencing
S1). 16S amplicon
demon- seq
Figure 3. Sequence-Base d Phylogenetic Composition of DTT-Treated Test Set 2 Samples after Standard and Benzonase Extraction
sample wasfrom
treated
each ofwith
sample thewas
mucolytic
treated agent
withCFthe
dithiothreitol
mucolytic
samplesagent
strated dithiothreitol
decreases in metagenomic
the
strated
relative
decreases
abundance
in theofrelative
P. aeruginosa
abundance
This issue may be particularly important after antibiotics, modulators
DNA
16S
extracted
(DTT) (Burnssequencing
amplicon and Rolain,
4 DTT-treated,
(DTT)
(16S 2014)(Burns
Amplicon,
homogenized
and then
and Rolain,
mixed
bottom row). to2014)
sputum
fullyand
homoge-
then mixed
was analyzed
after
using both
to benzonase1
fully homoge-and after
benzonase2
sequencing
benzonase1extractions
(MetaPhlAn2,
and benzonase2
top row) and
that were extractions
similar
 Some promising options for P. aeruginosa
detection
• Serology

• DNA-based:

• Species-specific PCR

• Sequencing

• Metabolites
 Exhaled breath condensate for P. aeruginosa
detection

Examined volatile chemicals exuded from bronchoalveolar lavage samples

• Mass spectrometry
Created panels of compounds to serve as “fingerprints” for different bacteria
Exhaled breath condensate for P. aeruginosa
detection

PPV 67% Compared

NPV 92% with culture
 Exhaled breath condensate for P. aeruginosa
detection
Advantages:
• Could be noninvasive

• Possible to perform on any age

• Possibility of fingerprints for subpopulations (multiresistant?)

Disadvantages:
• Equipment and expertise is not routinely available

• Very, very early in development

 Real-time diagnostics for Pseudomonas: Towards earlier
and better detection
Take-home points:
• Cultures are well-established and well-tested
• They have significant limitations
• Culture-independent techniques can overcome many of those
• Faster potential turnaround time
• Higher potential sensitivity, more comprehensive
• Limitations of culture-independent techniques include:
• Little to no clinical validation or standardization
• Expense, rare equipment and expertise
• Continued sampling issues, viability of cells is not always clear
• Molecular detection is now the state of the art for respiratory viruses
• MALDI and PCR have already streamlined P. aeruginosa speciation
• More study is needed to refine and validate better methods
Real-time diagnostics for Pseudomonas: Towards earlier
and better detection

See also:

Symposium S20: Modernising microbial diagnostics

Friday, 10:30
 Many thanks!
Dan Wolter John LiPuma
Chris Pope
Maria Nelson
Robyn Marsh
Michael Lee
Ken Bruce Anne Chang
Angshita Dutta
Geraint Rogers
Sam Miller
Fran Stressmann Barbara Kahl
Kyle Hager Jane Burns
Matthew Radey Margaret Rosenfeld
Anh Vo Ron Gibson
Eli Weiss Bonnie Ramsey
Mitchell Brittnacher Xuan Qin
Hillary Hayden Anne Marie Buccat
Elhanan Borenstein Michael Tunney
Alex Eng
Nicole Hamblett Dutch VanDevanter Our study subjects and
Pradeep Singh their families