1. Putting Drugs Where They Need
To Go
Azeem Sarwar

2. Why Does it Matter ?
Current practice of drug administration Systemic
(oral intake, intra vascular/muscular injections etc.)
Main problems associated:
•Even bio-distribution of pharmaceuticals throughout the body.
•Necessity of a large total dose of a drug to achieve high local concentration, with
adverse side-effects.
Less than 0.1% of the administered Tumor gets 1/30 99.9% goes to the
drugs are taken up by tumor cells Administered of this healthy tissues !!
in chemotherapy drug
• Drugs will go only where ever the blood goes – inner ear, brain, and eye are
behind a blood brain barrier and will not receive any drugs
Drug targeting or placing the drug where it needs to go (e.g. to tumor) can help:
•Increase the concentration of the drug at its targeted site improving drug efficacy.
•Lower the concentration at non-targeted site decreasing the toxic side effects.
•Decrease the total drug intake.
•Enable drug delivery across the blood barrier.
Can provide improved treatment and Help cure MORE people!

3. Magnetic Drug Targeting
(in general)

4. The Basics of Magnetic Drug Targeting
Magnetic drug delivery is Magnetic carriers can be drug coated
the transport or focusing nanoparticles
of therapeutic magnetic (e.g. sized to
carriers to regions of extravasate
disease by applied into tumors)
magnetic fields 100 –
Chemicell
250 nm Germany
Or they can be other magnetic things …
Polymer capsule Micelle with Cell with mag
with magnetic cores mag cores cores in or out
The first thing to ensure (obviously) is that the carriers are safe and bio-compatible.

5. The Basics of Magnetic Drug Targeting (continued)
The magnetic force
on an object scales
with its volume making the particle ×10 bigger increase the magnetic force by ×1,000
Magnetic fields
and forces fall
magnetic field strength |H| ~ 1/x3
off quickly with or
the distance away magnetic force |F| ~ 1/x7
from a magnet
inoperable x
now ON tumor
Magnetic fields are safe for people up to pretty strong fields (think MRI), but there is a
limit (first due to inducing eddy currents). United States FDA has set 8 Tesla (adults),
4 Tesla (children) magnetic field, and 20 Tesla/second rate of change, as safe limits.
Current state of the art:
Successful phase I inoperable 0.8 Tesla
human trials to focus tumor magnet
Control goal:
magnetized Do better
chemotherapy to shallow than a magnet
on a stick ...
inoperable tumors by a
single permanent magnet 0.8 Tesla

6. Forces on Magnetic Nanoparticles
In magnetic targeting, the force on a single particle can be written as
 1
( )
2
F= k∇ H where k is a constant and H is the magnetic field
2
This expression implies that a single particle
will experience a force from a magnetic field high low
H2 H2
region of low intensity to a magnetic field
region of high intensity
Thus any single magnet will attract and polarity doesn’t matter (flipping H to –H
particles to itself … leaves force, which goes as H2, the same)
nano
force F particle force F
N
S
N
S

7. Magnetic Drug Targeting to Ear

8. Why Ear?
• Ear disorder and disease affects millions of patients worldwide. The most common
reason for visiting a US doctor is ear infection.
• Tinnitus, commonly described as a ringing or roaring sound in the ears is
experienced by 1 in 10 people in the US.
• It can be very loud (90 dB) and can even cause suicidal tendencies.
• Noise-induced hearing loss , and Ménière’s disease can cause Tinnitus.
• No effective cure available – inner ear is behind blood-labyrinth barrier!
• Sudden Sensorineural Hearing Loss (SSHL) – now you hear tomorrow you won’t.
• Happens most often to people between the ages of 30 and 60
• 4000 new cases reported in the US every year.
• No effective cure available – inner ear is behind blood-labyrinth barrier!
• Middle Ear Infections (Otitis Media): Middle ear infections are most common in
children under 5 years of age.
• Treatments (antibiotics) exists for acute infections, but the success rate varies
and recurrences are common.
• Once acute infections develop into chronic, surgical procedure is required to
treat the infection.

9. Single magnets pull. But there are cases where it is good to magnetically
push (to inject).
Specifically ...
The Inner Ear is Behind the
Blood-Labyrinthine Barrier
(which is similar to the
blood-brain barrier)
Blood vessels that supply blood to most But blood vessels that supply blood to the
of the body have small pores. This allows inner ear have impermeable walls (to carefully
drug molecules to diffuse out from blood protect the inner ear, and brain and eyes, from
into surrounding tissue. If a patient eats a any potential contamination). This means that
pill or is injected with drugs, those drugs even though drugs are thought to exist for many
can get into most tissue. drug molecules
inner ear pathologies, they cannot reach the inner
can exit
ear.
BLOOD
Endothelial cells
There are a a number of inner ear diseases (tinnitus, sudden hearing loss,
Meniere’s) that affect millions of patients, where drugs exist (e.g. steroids) but it is
thought they do not reach the inner ear sufficiently to have therapeutic effect.

10. The developed magnetic push
treatment to reach the inner ear ...
Step 1: Inject gel with magnetic particles
into the middle ear through the ear drum.

11. The developed magnetic push
treatment to reach the inner ear ...
Step 2: Use the magnetic injector to non-
invasively push therapeutic nanoparticles
through the round window membrane into
the inner ear.
S
N
magnetic push force
S
N
magnetic
injector
It is much better to 3-5 cm
push over this short
3-5 cm distance Then to pull from the opposite
side of the head over 12-15 cm

12. It is possible/standard to locally deliver drugs to the middle ear
http://www.youtube.com/watch?v=FyZ0OLFEAm8
Magnetic push allows us to extend the delivery to reach the inner ear.

13. Pull Is Not Good Enough ...
Single magnets applied magnetic
pull nanoparticles force drops
towards them ... magnetic field
quickly with
or ||H|| ~ 1 / x3
distance from
magnet ||F|| ~ 1 / x7
x
now ON
Kopke et al pulled But even small
through the width humans have
11 T
of a rodents head big heads
~ 2 cm
> 10 cm
To generate the same force as created by Kopke in guinea pigs, but
over a 10 cm human head distance, would require a 11 Tesla magnet
For comparison, clinical MRIs are in the range of 1-4 Tesla, magnetic field bio-effects
can first be observed at 2 Tesla and higher (Allen 2001, Schenk 2005, Chakeres 2005).

14. How We Push: Create point of zero magnetic field at a distance (unstable node)
Create a point of zero magnetic field at a A single magnet
distance, a null or cancellation point.
Forces will emanate out from that null point.
N
S
Field lines (H) go from North to South
S
The Magnetic Injector System
N
Tilt, Flip, and H1 H≠0
Combine 2 Magnets (magnetic
fields do
Fields H just add,
+ not cancel)
S
Maxwell’s equations
N
are linear (H=H1+H2) = MIS
F
S
H2
N
S
N
H=0
(magnetic
fields
cancel)

15. at point C)
How We Pushed: Simulation (of Maxwell’s equations) + 1st Experiments
d) Magnetic Field e) Resulting Magnetic Forces
S
S
S
S
S
S
S
S
S
S
S
S
N
N
N
N
N
N
N
N
N
N
N
N
C
push region
S
S
N
N

16. Pushed 300 nm diameter MNPs into inner ear (cochlea) in rats
N
S S
N
And we have verified this treatment leaves rat hearing unharmed.

17. How Can We Do This Better ? (here comes math)
Goal: Maximize push force at point (x0, y0) in
Y
space.
How: Consider a matrix of magnets 1 2
(Halbach array). Need to figure out optimum (x0, y0)
θ X
magnetization direction (angle θ ) for each Pull Push
magnet.
N
Mathematical Formulation: Use analytical
expressions for simple bar magnet (Engel- Halbach Array
Herbert et al)
2
( )

A( x, y ) magnetic field eqn for Objective: min ∇ H (x0 , y0 )
 α i , βi x
B ( x, y ) magnetic field eqn for Subject to
α i2 + β i2 ≤ 1
Eqn for ith element magnetic field: Can be written in matrix from
   T  T 
H i ( x0 , y0 ) = α i Ai ( x0 , y0 ) + β i Bi ( x0 , y0 ) min q Pq : q Qi q ≤ 1,1 ≤ i ≤ N
q
Then for the entire Halbach array: 
for appropriate q , P, and Q.
 N  
H ( x0 , y0 ) = ∑ α i Ai ( x0 , y0 ) + β i Bi ( x0 , y0 ) P and Q are symmetric, but not semi-
i =1
definite. Non-Convex problem!

18. How Can We Do This Better ? (more math)
Step 1: Semi-definite Relaxation (SDR)
For symmetric matrices P, and Q
     
q T Pq = Tr(Pqq T ), q T Qi q = Tr(Qi qq T ) SDR optimizer

Define, G := qq T . Now the new problem is: Sub-optimal feasible solution
extracted from SDR optimizer
min Tr(PG) : Tr(QiG) ≤ 1,1 ≤ i ≤ N Convex
G
Upper bound
And Q ≥ 0, rank ( Q ) = 1 Non-convex
The relaxed problem is convex with global
optimum guaranteed. Q* is the optimizer.
Step 2: Majorization Method

The eigen vector q# corresponding to the
largest eigen value of Q* is a feasible solution.
 T  
The majorizer for f (q) := q Pq , at q# , is : Lower bound
        
F ( q ) : = f ( q # ) + ( q − q # ) T ∇ f ( q ) q + λ ( q − q # ) T ( q − q # )
  
#
   1 2 3 4
f (q) ≤ F(q) for q and at q# , F(q# ) = f (q# )
 Number of iterations
all ,
If q minimizes F(q), then we have
##
   
f (q## ) ≤ F(q## ) ≤ F(q# ) = f (q# )

19. How Can We Do This Better (design examples)
Pull Force Design: Generates
×5 more pull force than the
bench mark magnet (of same
Push Force Design: Generates volume)
×26 more push force than the
bench mark magnet (of same
volume)

20. Magnetic Pushing at Human Head Working Distance
A simple two magnet optimal design Force profile of the designed magnet
10
8× 10
5
6
Target region
10 10
8× 10 8× 10
4
(∇ H∇) 2) [A 3] ]
z [cm ]
3
0 6 6
( H [A /m/m
4
Push 4
(∇ H )x [A /m ]
3
2
2 Push Push
2
2
2
2
xx
0 0
2
-5 -2 0
2
-2
Pull Pull
-4 -4
-2
-6 -6
-5 Pull 2.5
3cm 3 3.5 4 4.5
5cm
5 5.5 6
Distance from the yz face of the magnet [cm]
0 5 -4
0
5 -5 -6
x [c m ] y [c m ] Magnet
magnetic injector held for 1 hour
scrape for push of red
at a 3 cm adult-human magnet-
fluorescent nanoparticles
to-inner-ear distance

21. Magnets can be nasty!

22. Animal Experiments

23. Treating Noise Trauma Hearing Loss
We are carrying out rat experiments for noise trauma hearing loss.
But how would you get a rat to tell if its hearing is good or bad?
We have a very nice rat If you startle a rat, it jumps
model to measure
hearing loss ... What
happened?!?
If rat has hearing
and if you first warn the rat, it jumps less high loss, it can’t hear Huh? What
the beep warning, happened?!?
jumps
Stop doing
higher
amplitude
amplitude
that!
BEEP
rat’s hearing loss
(e.g. at 10 kHz) time time

24. Noise Trauma Hearing Loss: Preliminary Rat Experiment Results
Induced hearing loss in rats by 1 hour of 1/3rd octave, 118 dB band of
noise centered on 16kHz.
Then we injected Chemicell 300 nm diameter particles coated with the
steroid prednisolone (an anti inflammatory) into their inner ears and
measured jump heights after a tone-warning-then-startle sequence ...
A rat with no magnetic treatment exhibited low A rat with magnetic treatment exhibited
pre-pulse inhibition of the startle reflex for a normal pre-pulse inhibition of the startle
tone warning in a 12-20 kHz range. They could reflex. They could hear the tone warning.
not hear the tones in this range.

25. We are also testing our magnetic treatment for noise-induced tinnitus ...
 Tinnitus is the perception of sound (ringing or roaring) when no sound is present.
 It is common: 1 in 10 people have noticeable level of tinnitus.
 It can be debilitating (up to 90 dB), in some cases leading to suicidal tendencies.
Currently, no effective treatments for tinnitus (Action On Hearing Loss market report, Goldman and Holme 2010)
We also have a very We already know that if you
nice rat model to startle a rat, it jumps
measure tinnitus ... What
happened?!?
If you first warn the rat, it jumps less high Can also warn the rat with a silent gap
I knew that
Stop doing was coming
amplitude
amplitude
that!
BEEP SILENT
GAP
constant tone (e.g. at 12 kHz)
time time

26. We are also testing our magnetic treatment for noise-induced tinnitus ...
 Tinnitus is the perception of sound (ringing or roaring) when no sound is present.
 It is common: 1 in 7 people have noticeable level of tinnitus.
 It can be debilitating (up to 90 dB), in some cases leading to suicidal tendencies.
Currently, no effective treatments for tinnitus (Action On Hearing Loss market report, Goldman and Holme 2010)
We also have a very We already know that if you
nice rat model to startle a rat, it jumps
measure tinnitus ... What
happened?!?
If rat has tinnitus
If you first warn the rat, it jumps less high Huh? What
It can’t hear the
happened?!?
gap warning,
Stop doing
jumps
amplitude
amplitude
that! higher
BEEP
constanttinnitus at 12 kHz)
rat’s tone (e.g.
time time

27. Induced tinnitus in rats by 1 hour of 114 dB noise, 1/3rd of an octave (114
dB is loud enough to create tinnitus but not hearing loss, measured by ABR).
Then we injected Chemicell 300 nm diameter particles coated with the
steroid prednisolone (an anti inflammatory) into their inner ears and
measured jump heights after a silent-warning-then-startle sequence ...
Rats with no magnetic treatment exhibited less Rats with magnetic treatment exhibited
pre-pulse inhibition of acoustic startle for a normal pre-pulse inhibition of acoustic
silent-gap in a 12-16 kHz background tone. They startle. They could hear the background
could not hear the tones under their tinnitus. tone and the silent gap warning.

28. Middle Ear Infections (Otitis Media)
• Single biggest reason to visit a doctor in the US – over 15 million cases in the US
reported last year
• Mostly occur in children under 5 years of age
• Current state of the art :
Step 1: Systemic ingestion of antibiotics
One millionth of the total blood flow reaches the middle ear and, hence, may
require strong drug dosage that only reduces the symptoms!
Bacteria in the middle ear may even develop immunity to the antibiotics
because of the low dosage that reaches the middle ear which is BAD!
Step 2: Puncture the ear drum and put a tube through it to deliver medication
Tympanostomy Doesn’t
tube to access make kids
the middle ear happy
(Dohar 2011)

29. Middle Ear: Treating Otitis Media without Ear Drum Puncture
(Preliminary results) Investigating if magnetic injection can deliver drugs to the middle
ear, without a need for ear drum puncture magnetic
magnetic
particles
injector
placed in
outer ear
vs
S
N
Tympanostomy
S
tube to access
N
magnetic
the middle ear transport
(Dohar 2011)
First Test: Rat Middle Ear Tissue Scrapes
no magnetic force, no with magnetic force
observed particles at the observed fluorescent particles
back of the middle ear on cochlea bone (red speckles)
vs
200 µm 200 µm

30. The Big Picture!
Magnetic
Drug Targeting
Clinical
Controls
Practice
Nano Science

31. The Big Picture!
Magnetic
Drug Targeting
Clinical
Controls
Practice
Drug Targeting in Lungs Directing Therapy to Strokes
Nano Science
Inhaled
therapeutic
Tumor
particles
Medication for stroke can cause blood
Magnetic field leakage - Need to steer the drug at the
shaping clot location!

32. Acknowledgements
Collaborators
Benjamin Shapiro – UMD
Didier Depireux – UMD
Roger Lee – UMD
Diego Preciado – Children’s National Medical Center, DC
Arkadi Nemirovski – Georgia Tech
Undergraduates
Reza Basiri – UMD
Mohammad Ahmed– UMD
Funding support
Children’s National Institute for Pediatric Surgical Innovation;
State of Maryland TEDCO (Technology Development Corporation)
and MIPS (Maryland Industrial Partnerships); Action on Hearing
Loss and NIGMS-NIH (National Institutes of Health).

33. Thank you for your attention!