Use of High-Throughput Tools for Telescoped Continuous Flow Synthesis of an Alkynylnaphthyridine Anticancer Agent, HSN608

Article abstract of DOI:10.1021/acs.oprd.0c00289

Developing continuous syntheses of lead compounds to support in vivo studies and preclinical evaluation remains an underdeveloped area. We report a telescoped continuous flow synthesis of an alkynylnaphthyridine lead compound for the treatment of FLT3 mutations in acute myeloid leukemia. Different strategies were used to develop the route, including Design of Experiments (DoE), high-throughput experimentation (HTE), and application of desorption electrospray ionization mass spectrometry (DESI-MS) to optimize and telescope the amidation and Sonogashira couplings to prepare the target compound, HSN608, a potent FLT3 inhibitor. Findings from these statistical design and automation studies helped streamline our workflow to achieve 10-fold and 5-fold reductions in the catalyst and cocatalyst loadings, respectively, in the synthesis. The application of high-throughput tools combined with a telescoped continuous synthesis method enabled an efficient and safe synthesis of this lead compound using the hazardous coupling reagent HATU while minimizing byproduct formation.

Full text of DOI:10.1021/acs.oprd.0c00289

pubs.acs.org/OPRD
Article
Use of High-Throughput Tools for Telescoped Continuous Flow
Synthesis of an Alkynylnaphthyridine Anticancer Agent, HSN608
Shruti A. Biyani,† Qingqing Qi,† Jingze Wu, Yuta Moriuchi, Elizabeth A. Larocque, Herman O. Sintim,
and David H. Thompson*
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00289
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Developing continuous syntheses of lead compounds to support in vivo studies and preclinical evaluation remains an
underdeveloped area. We report a telescoped continuous flow synthesis of an alkynylnaphthyridine lead compound for the treatment
of FLT3 mutations in acute myeloid leukemia. Different strategies were used to develop the route, including Design of Experiments
(DoE), high-throughput experimentation (HTE), and application of desorption electrospray ionization mass spectrometry (DESI-
MS) to optimize and telescope the amidation and Sonogashira couplings to prepare the target compound, HSN608, a potent FLT3
inhibitor. Findings from these statistical design and automation studies helped streamline our workflow to achieve 10-fold and 5-fold
reductions in the catalyst and cocatalyst loadings, respectively, in the synthesis. The application of high-throughput tools combined
with a telescoped continuous synthesis method enabled an efficient and safe synthesis of this lead compound using the hazardous
coupling reagent HATU while minimizing byproduct formation.
KEYWORDS: high-throughput experiment, continuous flow synthesis, Sonogashira coupling, Design of Experiments, FLT-3 inhibitor
10 mol % Pd and 5 mol % Cu, respectively, low yield, use of an
explosive amidation reagent (HATU)8 that would be unsafe
for large-scale batch synthesis, and a long reaction time of ∼24
h for the two steps9 (Figure 1). We sought to improve the
safety, efficiency, and heat and mass transfer profiles of this
synthesis10 by executing it in flow.11 Furthermore, continuous
flow methods also provide a rapid reaction optimization
platform12 to enable reductions in catalyst loadings that
otherwise contribute to high production costs and Pd-
associated toxicity13 in active pharmaceutical ingredients.
Reduced batch variability, smaller reactor footprint, and
minimized solvent waste would also render the overall process
greener.14 The aim of this study was to identify the optimal
conditions for the synthesis of HSN608 utilizing microfluidics
with the assistance of HTE tools. Once the optimized synthesis
was identified, it could be further scaled up to support
preclinical studies of this lead agent.
The one variable at a time (OVAT) approach provides only
local knowledge of reactivity patterns and is therefore blind to
the interactions between variables. The DoE approach is not
subject to this limitation, making it ideal for studying
interrelated variables.15 HTE utilizing a robotic liquid handling
system provides a platform to integrate automation with
synthesis that facilitates simultaneous interrogation of a wide
array of reaction condition variables simultaneously.16 While
INTRODUCTION
■
In spite of the growing interest in continuous synthesis of new
chemical entities by the pharmaceutical industry, the
application of Design of Experiments (DoE) strategies with
guidance from high-throughput experimentation (HTE) for
reaction optimization remains an underdeveloped area.1 Rapid
discovery of the most favorable and benign conditions for
integration into continuous synthesis schemes can also enable
the development of efficient and robust synthetic routes,
potentially reducing the time required to translate a medicinal
chemistry lead through clinical trials.2 One such medicinal
chemistry challenge is acute myeloid leukemia (AML), an
aggressive form of blood cancer that remains difficult to treat
despite tremendous efforts by academics and pharmaceutical
companies to find a durable cure.3 The 5-year survival rate for
AML is approximately 30% and drops below 10% in patients
over 65 years of age,4 highlighting the dire need for a better
treatment. Fms-like tyrosine kinase 3 (FLT-3) is a class of
receptor tyrosine kinase that is overexpressed in AML
patients.5 Nearly one-third of initially responding patients
have shown relapse over time by developing secondary
mutations during treatment.6 Sintim and co-workers discov-
ered HSN608 (5) (Scheme 1) as a lead compound with high
potency toward inhibition of all FLT-3 forms.6a,7,9 This lead
also displays inhibition activity toward RET and its mutant
forms better than the reported RET inhibitors.7 In order to
support the drug development effort to enable in vivo
experiments and toxicology studies for this lead compound,
identification of a synthetic route that can be rapidly upscaled
is essential.
Special Issue: Flow Chemistry Enabling Efficient
Synthesis
Received: June 20, 2020
The reported batch synthesis of HSN608 had several
drawbacks, including high catalyst and cocatalyst loadings of
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
a
Scheme 1. Reaction Sequence for the Synthesis of 5
a
For the amide coupling reaction (I), the optimal residence time and temperature were determined using 22 full-factorial design. Optimal
conditions for the Sonogashira coupling (II) were identified using DoE and HTE coupled to DESI-MS along with targeted batch experiments.
Figure 1. Conventional batch reaction method vs our approach utilizing DoE/HTE/DESI-MS → flow strategy for rapid discovery of favorable
conditions for the flow synthesis of compound 5. HTE using DESI-MS readout to guide DoE in batch and flow also provides mechanistic
information. Taken together, these data enable the rapid development of efficient synthesis conditions.
Figure 2. Microfluidic synthesis of 3 via HATU-mediated amide coupling in a glass reactor chip (staggered oriented ridge (SOR) mixer chip 3227
by Chemtrix). Conditions: A, 0.277 M 1 and 2 (1 equiv) and DIPEA (3 equiv) in DMF; B, HATU in DMF (1.1 equiv); ambient pressure; reactor
volume, 19.5 μL.
conventional LC−MS takes up to 33 h for analysis of 384
reactions, desorption electrospray ionization mass spectrome-
try (DESI-MS) is capable of analyzing up to 6144 reactions on
a single plate array in less than 2.5 h without any requirement
for workup. The large data set of full mass spectra at each point
in the array that is generated by this method enables rapid
searches for the m/z of interest and follow-up MS/MS
experiments using automated Chemical Reaction Integrated
Screening software (CHRIS) for target and/or byproduct
structure confirmation.17 Streamlining of our workflow in this
manner allowed us to quickly identify crucial reaction
parameters that informed our understanding of the reaction
B
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
mechanism in the key C−C bond-forming step. Using these
powerful new reaction discovery tools (Figure 1), we report a
strategy for expediting the optimization of transformations
involved in the two-step synthesis of a target molecule in flow
using DoE, HTE, and DESI-MS for identification of a
continuous flow synthetic route. Given the importance of
cross-coupling reactions in medicinal chemistry,18 execution of
the Sonogashira reaction to produce 5 is the key step in the
process. Although Sonogashira transformations in flow have
been reported,11,19 the heteroatom density in our target
molecule makes the flow synthesis of 5 challenging compared
with other reported targets.
RESULTS AND DISCUSSION
Figure 3. Contour plot of residence time, temperature, and isolated
yield from the HATU-mediated amide coupling using a 22 full-
factorial DoE evaluation. The yields of the reaction are color-coded,
where dark red indicates the highest yield regime and dark blue
indicates the lowest yield regime.
■
DoE for Amide Coupling in Flow. Our initial batch
experiments employed a number of common amide coupling
reagents (Table S1), but these typically resulted in poor yields
(<40%). The highest isolated yields of ∼55−60% were
observed in small-scale batch reactions with HATU. Since
reaction scale-up with this explosive reagent would be a safety
concern,8 we pursued a safer alternative by developing a
HATU-mediated amide coupling process in flow. Although the
use of HATU in flow reactions has been reported for peptide
synthesis,20 we sought to quickly identify optimal conditions
for its use in the preparation of intermediate 3.
Discovering Conditions for Sonogashira Coupling
Using Automation (DoE, HTE, and DESI-MS Experi-
ments). On the basis of reports of other Pd-catalyzed
reactions, we expected that the order of reagent addition
could be an important parameter for efficient execution of the
Sonogashira transformation, particularly when a a robotic
open-air liquid handling system is used.23 Our first set of
experiments focused on identifying the best order of addition
and an initial screen of catalysts. We performed this experiment
on the model substrates 5-bromoisoquinoline (6) and 3-
ethynylpyridine (8) (Scheme 2) for four different orders of
We chose DMF as the solvent since it is a low-volatility,
high-boiling polar aprotic solvent with a high dielectric
constant, although on scale it does have reproductive toxicity
as an exposure concern.21 Our initial flow experiments did not
use an in-line mixer (without T-mixer in Figure 2). HATU in
DMF was injected via one syringe port, and a 1 + 2 + DIPEA
mixture in DMF was injected directly at 50 °C into a glass
reactor chip via the second port. Very low product formation
was observed because of the formation of guanidino
derivatives, as observed by (+) ESI-MS (Figure S1).22 We
then changed the setup and employed a microstatic mixer prior
to the chip, which we hypothesized would allow for the
formation of activated ester at 20 °C using 0.378 M solutions
of the reagents. However, clogging due to the relatively low
solubility of HATU (1.1 equiv) in DMF was observed. Next,
we incorporated a T-mixer (Figure 2) at 20 °C to allow for the
formation of the activated ester before all of the reactants
entered the glass reactor chip held at 50 °C using reagent
concentrations of 0.277 M. After confirming the formation of
intermediate 3 by (+) ESI-MS (Figure S2) and observing the
product profile in the crude mixture by LC−MS (Figure S3),
we decided to quickly identify the most favorable conditions
using a 22 full-factorial design. We hypothesized that this
approach would allow us to identify high-yielding conditions in
a minimum number of experiments. Reaction times of 10 and
20 min and temperatures of 50 and 90 °C were chosen (Table
S2). Figure 3 summarizes the yield contours for 3 obtained
from the 22 full-factorial experiment.
Scheme 2. Model Substrates for Preliminary Rounds of
HTE for Sonogashira Coupling
addition and two different catalysts, Xphos Pd G3 and
Pd(OAc)2, at 70 °C in ACN (Figure S6). A total of 96
unique reactions with four replicates were performed via HTE
and analyzed using (+) ESI-MS, running a script in Perl
software to generate Excel data files as previously reported.23a
We found that Xphos Pd G3 gave the greatest number of hits.
These data also revealed that extended contact of the base and
catalyst at room temperature during liquid transfer without the
reactants present did not yield the desired product, likely as a
result of catalyst degradation since Pd(0) has lower stability in
the presence of airthe conditions that existed while the
reaction mixtures were being built by the liquid handling robot
(Figure S8).
In the second HTE campaign, our goal was to identify the
most crucial continuous variables and to understand the
interactions between the variables. We evaluated 5-bromo-8-
aminoisoquinoline (7) and 8 as model reactants (Scheme 2)
because of their structural similarity to 4 and 3, respectively.
We chose a 23 full-factorial DoE approach coupled with a
DESI-MS readout, where time, temperature, and stoichiometry
were chosen as the main HTE factors. The low and high levels
A midpoint experiment at 75 °C for 15 min was also
investigated, and 65% conversion was observed in that case
(Figure S5). Since isolation of the amide-coupled product by
column chromatography was tedious, we developed a
precipitation method for reaction workup wherein addition
of water at the end of the reaction produces a solid form of 3
that can be easily recovered by filtration in ∼86% yield (50 °C,
RT = 20 min). A further upscaling of this method to a simple
coiled tube reactor in a 50 °C oil bath enabled the rapid
synthesis of intermediate 3 (2.260 g total) at a rate of 96 mg/h.
C
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 4. HTE heat map for the Sonogashira coupling indicating product ion counts in each cell (means of four reaction replicates, where the
number corresponds to the MS ion count from DESI-MS). Levels: time = 90 and 240 min (indicated in light pink); temperature = 100 and 150 °C
(dark pink); five bases at 1.5 equiv and TEA at 1.5 and 3 equiv (aqua green); six solvents (light blue); stoichiometry = 1:1 and 1:2.5 (blue). The
scale based on the product ion count is shown using the color in the arrow at the bottom right, where dark red represents the highest product ion
count (more efficient or successful reaction) and white represents a lower product ion count (unsuccessful reaction). An ion count of 100 was
considered as the threshold, and cells with >100 counts were considered as successful reactions.
were time = 90 and 240 min, temperature = 100 and 150 °C,
and stoichiometry = 1:1 and 1:2.5 (with 8 in excess). This DoE
study was performed using six different solvents (DMF, NMP,
DMSO, nitromethane, DMAc, and EtOH) and five different
bases (K2CO3, K3PO4, piperidine, TEA, and DIPEA) to
produce 336 unique reactions on a single DESI-MS plate
(Figure 4). We had also transferred 30 μL of the reaction
mixture at 20 °C after all of the reagents were mixed in a 96-
well block before transfer in quadruplicate to a 384-well plate
for pinning onto a PTFE DESI plate at 90 and 240 min. The
product ion counts from the DESI-MS analysis were taken as
the output to make the interaction plots (Figure 5), identify
the most crucial variable, perform analysis of variance
(ANOVA), and develop Pareto plots (Figure S11).
We observed significant hits with solvent/base combinations
as indicated in the heat map in Figure 4, including EtOH/TEA,
NMP/K2CO3, NMP/K3PO4, and DMAc/piperidine. All of the
reaction mixtures were pinned onto the same PTFE plate for
DESI-MS analysis. We observed that temperature had the
greatest impact on the reaction efficiency (Figure 5) with lower
temperature (100 °C) particularly favoring higher product ion
intensities. A higher reaction temperature (150 °C) likely
results in catalyst deactivation since palladium catalyst
oxidation at elevated temperature has been previously reported
in the literature.24 The second most important parameter
revealed is the interaction between time and temperature,
while stoichiometry had little to no impact on the reaction
efficiency. We concluded from Figure 5 and the Pareto plot
(Figure S11) that the temperature, the time, and their
interaction were the three most crucial variables. Thus, we
then generated a contour plot that considered only these three
variables and along with the ion counts that we obtained at 20
°C (Figure 6) so that we could identify the most favorable
temperature for Sonogashira coupling. Using the contour plot
based on the DESI-MS product ion counts, we found the
optimal temperature range to be 78−103 °C, with reaction
initiation occurring at ≥55 °C (Figure 6).
After performing these high-throughput experiments with
the model substrates, we decided to move toward the
continuous flow synthesis of the model substrates followed
by the synthesis of HSN608 directly in flow. We performed the
continuous flow synthesis with the model substrates using a
glass microchip flow reactor mounted on a Labtrix S1 system
(Chemtrix BV, Echt, The Netherlands) and monitored
product formation by (+) ESI-MS. Having observed the
Sonogashira coupling product for the model substrates at the
respective m/z values from the flow synthesis (Figure S20), we
planned to directly translate these findings to HSN608 and
perform the optimization in flow. We investigated the flow
synthesis of 5 using 3, 4, and K2CO3 in NMP, as these gave us
hits in the HTE (Figure 4) and had also worked for the model
substrates. However, we did not see product 5 by (+) ESI-MS
even though the coupling product was observed for the model
compounds in flow. The poor solubility of 4 in EtOH also
limited our options in microfluidics. Therefore, we pursued
another round of HTE to identify the discrete variable most
favorable for the synthesis of 5.
On the basis of the findings from the previous DoE showing
that the reaction becomes kinetically favorable at ≥55 °C
(Figure 6), we performed the next HTE at 55 °C. The aim of
this round was to identify the most favorable solvent, base, and
ligand (section S8 in the Supporting Information (SI)). We
chose seven different bases/salts with pKa values ranging
between 10.25 to 11.40: TMEDA, TEA, TBAPF6, piperidine,
DIPEA, K2CO3, 2,2,6,6-tetramethylpiperidine (TMPH), and
we also selected [(t-Bu)3PH]BF4 as the ligand. We limited our
D
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 6. Contour plot of product ion counts vs time and temperature
to identify the optimal reaction time and temperature. The mean
product ion count values of four replicates for the case of EtOH as the
solvent and TEA as the base was used. The color-coded scale is used
for the product ion count as the response from DESI-MS. Dark red
represents the highest ion count (most optimal reaction conditions),
and blue represents the lowest ion count (least favorable reaction
conditions). The plot was generated using the statistical software
Minitab with X = temperature, Y = time, and Response = product ion
count. For 20 °C, the reaction mixture was pinned directly at 90 and
240 min without any heating. (Explanation of the plotting is
highlighted in Figure S12).
conditions are depicted in Figures S15−S17. Since base is
crucial for deprotonation to form the Cu−acetylide inter-
mediate (Scheme 3), we inferred from these findings (Figure
8) that the formation of this intermediate and thus cycle B is a
key step in the reaction and that enhancement of the
transmetalation step would likely lead to further increases in
the reaction efficiency.
Translation of the HTE Findings to Better Understand
the Mechanism for Coupling of Compounds 3 and 4 in
Batch and Flow. In order to test our conclusion that cycle B
is crucial as inferred from our HTE results, we performed two
sets of experiments in batch mode (Table 1) in which we
monitored the completion of the reaction by TLC at different
time points. Reaction completion was taken in this initial study
as disappearance of 3 by TLC analysis. We used 5-bromo-8-
amino-1,7-naphthyridine (4) and 5-iodo-8-amino-1,7-naph-
thyridine (9) in the Sonogashira coupling with different
catalysts and bases. We found that the reaction was complete
after 12 h with aryl bromide 4 but was not complete even after
15 h with aryl iodide 9. We interpret these findings to mean
that the two cycles are not well-synchronized with aryl iodide
9, with cycle A proceeding faster than cycle B, leading to an
increase in the decomposition products from cycle A, as
evidenced by more spots on TLC and faster consumption of 9.
In contrast, when the corresponding aryl bromide was used,
the two cycles A and B were better synchronized, resulting in
more of the desired product.
Figure 5. Interaction plots involving time (90 min, 240 min),
temperature (100 °C, 150 °C), and stoichiometry (1:1, 1:2.5). The
responses are indicated on the y axes as the DESI-MS product ion
counts obtained from the 23 full factorial design experiment. The
legends corresponding to the specific conditions are depicted below
the individual graphs. (A) The blue line represents 90 min, and the
orange line represents 240 min. (B, C) The blue line represents 1:1
stoichiometry, and the orange line represents 1:2.5 stoichiometry. In
cases of 1:2.5 stoichiometry, alkyne 8 was used in excess.
experiment to three solvents (DMF, NMP, and DMAc)
because of the poor solubility of 4 in other solvents. We also
added RuPhos and BrettPhos to the HTE since these ligands
reportedly enhance oxidative addition and reductive elimi-
nation.25 The overall results are shown in Figure 7. Reactions
in DMF were found to provide the widest choice of bases
compared with NMP and DMAc. Interestingly, reactions with
no additional ligand showed higher reaction efficiency than
those loaded with RuPhos or BrettPhos (bigger circle size for
green as indicated in Figure 7). The more favorable bases were
the ones with higher pKa values, such as piperidine, TMPH,
and salts like K2CO3, TBAPF6, and the ligand [(t-Bu)3PH]BF4
(Figures 7 and 8). Bar graphs with individual solvent
Since oxidative addition rates are known to follow R−Br <
R−I,26 we concluded that this step in the Sonogashira reaction
is not the rate-limiting step in the formation of 5. Further, we
compared Cu-free reactions27 and reactions with 3% CuI (SI
section S9a) and found that it takes >4 h to complete the
reaction without Cu(I) but that the reaction was complete in 1
h in the presence of Cu(I). We inferred from these studies and
E
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 7. Global analysis and trends for 288 unique Sonogashira reactions at 55 °C from HTE using 3 and 4 as substrates at m/z 546.2. Xphos Pd
G3 was used as the catalyst, with average product ion counts calculated from ion counts measured for reactions that were pinned onto the DESI
plate in duplicate. The three grids refer to the three solvents DMAC, DMF, and NMP, and the fill colors represent the ligand condition (pink =
BrettPhos, blue = RuPhos, green = no additional ligand). Each column point represents one of four different time points (0 h is when the solution
from the 384-well plate was pinned immediately onto the PTFE plate, while the other time points are 1.5, 6.75, and 12 h at 55 °C). Each row of
conditions corresponds to an additive: seven bases and one ligand ([P(t-Bu)3H]BF4). The size of the bubble corresponds to the average product
ion count from DESI-MS, with larger-diameter spots representing higher ion counts and thus more efficient reaction conditions.
Scheme 3. HTE DESI-MS and Batch Experiments
Corroborated That Cycle B Is Crucial and Oxidative
Addition Is Not the Rate-Limiting Step in the Sonogashira
Reaction to Form 5 from 3 and 4
Figure 8. DESI-MS HTE plot for the Sonogashira coupling for the
synthesis of HSN608. The reaction conditions were Xphos Pd G3,
DMF, 1.5 h, 55 °C. The structures of the seven bases and one ligand
are shown. Additional ligand conditions are color-coded for each type
Table 1. Sonogashira Time-to-Completion Experiments
with Alkyne 3 in Batch Mode
a
(pink = BrettPhos, green = no ligand, purple = RuPhos). The red
time
b
dashed line represents a signal-to-noise ratio of 3 as the cutoff. Noise
signals were measured by taking the average at a region on the PTFE
plate where no reaction mixture was pinned. Product ion (m/z 546.2,
[5 + H]+) counts above this cutoff are considered to indicate
successful reactions, with higher ion counts corresponding to more
efficient reaction conditions.
entry halide
catalyst
CuI?
base
(h)
c
d
d
1
2
3
4
4
9
4
4
PdCl2(PPh3)2/XPhos
PdCl2(PPh3)2/XPhos
XPhos Pd G3
[PdCl(allyl)]2/
[(t-Bu)3PH]BF4
yes
yes
yes
no
TMPH
TMPH
piperidine
piperidine
12
>15
1
>4
5
4
Pd(amphos)Cl2
no
Cs2CO3
>4
a
b
in agreement with literature28 that the transmetalation process
was likely the slow step in the reaction and required further
optimization (Scheme 3).
The reactions were run at 80 °C unless otherwise mentioned. The
c
time for complete consumption of 3 was monitored by TLC. TMPH
= 2,2,6,6 tetramethylpiperidine. The reaction was run at 55 °C.
d
We then adapted these initial findings to flow, where the first
challenge was to address the relatively poor solubility of 4
(14.6 mg/mL, 0.065 M at 20 °C in DMF). Using a syringe
heater in the reactor setup solved the problem (solubility of 4:
38.1 mg/mL, 0.170 M at 55 °C in DMF), but as it would
require an additional heating device for the tubing and syringe,
we focused on 4 at 20 °C by recrystallizing it twice, first from
EtOH and then from EtOAc. This treatment enabled a
solubility increase to 0.0785 M (17.6 mg/mL in DMF). The
F
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
reaction setup used a T-mixer at 20 °C prior to injection of the
solution of 4 into the heated glass reactor chip that was held at
80 or 95 °C (Figure S21). Using TLC and LC−MS for
reaction analysis, we found that the times required for
complete conversion of 3 in the batch and flow trans-
formations were the same (∼60 min). We did not attempt to
accelerate the reaction rate by increasing the temperature
above 100 °C since this would lead to catalyst degradation as
suggested by Figure 6, with subsequent reactor chip clogging.
On the basis of these findings, we sought to modify our
strategy for execution of this reaction in flow.
Synchronizing the Two Cycles to Enhance the
Reaction Efficiency. Given our findings that the trans-
metalation rate needed acceleration, we sought to increase the
synchronization of cycles A and B in Scheme 2 by enhancing
both the oxidative addition and Cu−acetylide formation steps
simultaneously. To enhance the oxidative addition, we used
the electron-donating ligand [(t-Bu)3PH]BF4 and catalyst
systems like PdCl2(PPh3)2 and PdCl2(MeCN)2, as shown in
Table 2. Furthermore, pyrrolidine was chosen as the base
Reactions run at 0.075 M 4 caused deposition of solid in the
syringe over time, stalling the pump within 8 h (Table 3, entry
1). Lowering the concentration of 4 to 0.065 M overcame this
problem. We then focused our effort on determining the
optimal residence time for the second step in the telescoped
reaction (RT2). This optimization was achieved by connecting
CFIs29 in series (Figure S27), producing an overall yield of
54% when lower flow rates and two CFI reactors in series were
used (entry 4). We learned that for the telescoped flow
reaction synthesis, lowering the catalyst loadings from 1 to 0.5
mol % can reduce the yield from 54% to 32% (entries 4 and 5).
Furthermore, an optimal residence time, achieved by
connecting the appropriate number of CFIs in series, is crucial
for an efficient telescoped flow synthesis. A plausible
explanation for this is that with the lower residence time
resulting from fewer CFIs (entry 2), complete conversion of
alkyne 3 cannot be achieved, producing lower yields.
Conversely, a larger number of CFIs (entry 3) can possibly
introduce oxygen infusion from the surroundings between the
connections, contributing to the formation of the homocou-
pling byproduct and reducing the yield of the desired product.
By balancing these factors, we have identified a telescoped flow
synthesis route for synthesis of 5 at 8 mg/h with continuous
operation for >55 h to enable the telescoped synthesis of
HSN608.
Overall, we have shown that the use of high-throughput
tools and a Design of Experiments approach along with
targeted bath mechanistic experiments were able to rapidly
guide the optimization and identification of a telescoped
continuous flow synthesis of the potent FLT-3 inhibitor
HSN608. Scale-up of this route will be crucial to further
evaluate the in vivo efficacy and toxicology of this lead
compound. The capabilities of these tools can also be
leveraged to develop efficient syntheses of other lead
compounds in order to support drug discovery efforts. The
flow methodology developed in this study can also be used for
synthesis of other kinase inhibitors with similar structural
motifs, like ponatinib.30
Table 2. Catalyst, Ligand, and Base Selection for the
Sonogashira Coupling of 3 and 4 in Flow Using LC−MS to
a
Monitor the Product Yield
temp.
(°C)
entry
catalyst/ligand
base
yield (%)
1
2
3
PdCl2/[(t-Bu)3PH]BF4
PdCl2/[(t-Bu)3PH]BF4
80
95
80
pyrrolidine
pyrrolidine
Cs2CO3
10
49
27
PdCl2(MeCN)2/
b
[(t-Bu)3PH]BF4
4
PdCl2(MeCN)2/
80
pyrrolidine
88
56
c
[(t-Bu)3PH]BF4
5
6
7
PdCl2(PPh3)2
80
95
95
pyrrolidine
pyrrolidine
pyrrolidine
d
e
PdCl2(PPh3)2
85 (79 )
f
g
PdCl2(PPh3)2
73
a
Conditions: DMF as the solvent, 3% Pd catalyst, 3% [(t-
Bu)3PH]BF4, 3% CuI, RT = 30 min, unless otherwise noted.
b
c
d
DMF/H2O. 6% catalyst, 4% CuI, 6% ligand. 1% catalyst, 1%
e
f
CuI. The isolated yield is given in parentheses. 0.5% catalyst, 0.5%
CuI. The time for >90% conversion of 3 was ∼80 min under batch
conditions.
g
CONCLUSIONS
■
We have shown that a 22 full-factorial design can be used to
quickly optimize HATU-mediated amide coupling and enable
safer handling of this explosive reagent on scale using a
microfluidic flow reactor. A simple precipitation workup was
also developed that circumvents the use of column
chromatography for the purification of 3. This work has also
demonstrated that a strategy employing a Design of Experi-
ments/high-throughput experimentation approach coupled
with desorption electrospray ionization mass spectrometry
can rapidly identify the most favorable batch and flow
conditions for Sonogashira coupling. These studies revealed
the optimal choice of base, solvent, and temperature conditions
as well as identifying that cycle B is a key rate-limiting step in
the formation of 5. These findings focused our workflow
toward synchronizing the two cycles through catalyst and base
screening under flow conditions to enhance the reaction
efficiency. We then applied these findings to the identification
of a continuous flow synthesis route for 5, a clinical candidate
for the potential for treatment of acute myeloid leukemia. The
catalyst and cocatalyst loadings were reduced by 10-fold and 5-
fold, respectively, along with improvements in the overall yield.
These findings show that DoE coupled with HTE DESI-MS
provides an opportunity for rapid optimization of synthetic
because of its relatively high pKa of 11.27 to enhance Cu−
acetylide formation, as higher ion counts were observed in
HTE with bases having higher pKa values (piperidine and
TMPH, Table 2). K2CO3 was not chosen because of its poor
solubility in DMF. We observed an increase in yield with
increasing temperature, consistent with our DoE findings
(entry 6). LC−MS analysis of the crude reaction mixtures
(Table 2) revealed lower amounts of byproduct formation
using PdCl2(PPh3)2 as the catalyst compared with PdCl2/[(t-
Bu)3PH]BF4, with >90% conversion of 3 seen in 30 min
(Figure S22). Lowering the catalyst loadings to 0.5% increased
the time required for the reaction more than 2-fold. Thus, we
chose 1% PdCl2(PPh3)2/pyrrolidine in DMF at 95 °C as our
optimal conditions for the telescoped flow synthesis of
HSN608.
Telescoped Continuous Flow Synthesis of HSN608.
For our telescoped flow synthesis efforts, we built a modular
flow reactor comprising a four-way cross-assembly, two
Harvard syringe pumps mounted with 25 mL syringes, and
coiled flow inverter (CFI) reactors made from PFA tubing, as
shown in Figure 9.
G
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 9. Telescoped continuous flow synthesis of HSN608. f1, f 2, f 3, and f4 are the flow rates. The residence time for the amide coupling was 20
min at 50 °C. The residence time for the Sonogashira coupling (RT2) was 2, 4, or 6 h at 95 °C.
of the reaction mixtures. The heating blocks were brought to
room temperature prior to pinning onto the PTFE-coated
DESI plate using the pin tool. The pinning method described
previously for transfer of 50 nL volumes onto the porous PTFE
surface using the pin tool was employed.17
Table 3. Flow Reactor Conditions Evaluated in the
Telescoped Synthesis of HSN608 Using the Reactor
Configuration Shown in Figure 9
a
b
c
entry
f1, f2, f 3, f4 (μL/min)
RT2 (h) no. of CFIs
yield (%)
d
1
2
3
4
5
14.26, 4.326, 4.45, 4.45
14.26, 4.326, 4.45, 4.45
14.26, 4.326, 4.45, 4.45
7.13, 2.16, 2.225, 2.225
7.13, 2.16, 2.225, 2.225
2
2
6
4
4
1
1
3
2
2
pump stalled
Mass Spectrometry. ESI-MS (to identify the order of
addition in HTE and determine the yield from continuous flow
synthesis) was performed using a Thermo Fisher TSQ
Quantum Access MAX mass spectrometer. This tool was
connected to a Dionex Ultimate 3000 series pump and WPS-
3000 autosampler. An autosampler capable of handling 96-well
plates was used for the analysis with a vapor-tight seal on the
96-well plate. Data recording was achieved using parameters
optimized for the ESI source, wherein the spray solvent was
ACN with 0.1% formic acid or MeOH with 0.1% formic acid,
the spray voltage was +5 kV (positive mode), the capillary
temperature was 250 °C, and the injection volume was 1−5
μL. For the analysis and processing of data, Thermo Fisher
Xcalibur software was utilized.
15
30
54
32
e
a
Conditions: [4] = 0.065 M unless otherwise noted. The residence
time for amide coupling (RT1) was 20 min. In the second step, 1%
b
PdCl2(PPh3)2 and 1% CuI were used. Number of CFIs connected in
c
d
e
series. Isolated overall yields. [4] = [0.075 M]. 0.5% PdCl2(PPh3)2
and 0.5% CuI. The reaction was run at 95 °C.
transformations compared with conventional trial-and-error
batch methods. Efforts are now underway to scale up the
continuous synthesis of 5.
DESI-MS analysis was performed using a previously
published method,17 except that samples at a density of
3072/plate and 1536/plate were used instead of the reported
6144/plate. In order to collect the DESI-MS data, we used a
linear ion trap mass spectrometer (LTQ XL, Thermo
Scientific) along with a DESI imaging source (DESI 2D,
Prosolia Inc.). Data acquisition was done using the Xcalibur
software, version 3.0. The spray solvent used was MeOH or
MeOH with 0.1% formic acid. The data generated were
collected in the form of “yes/no” output for each spot using in-
house-built software to generate a heat map and Excel files.
The average product ion counts from the Excel file generated
for the Sonogashira coupling were used for the DoE analysis in
Minitab and for other analyses using either Excel or R
programming. Minitab is a user-friendly statistical software that
helps with performing analysis and evaluation of statistics on
data provided as input. It also helps in the identification of
trends and patterns to decipher and extrapolate solutions to a
data set containing problem. The response or Excel sheets
could be directly used in the Minitab, enabling the investigator
to perform ANOVA and generate contour plots and Pareto
charts after choosing the input, output, and response.
EXPERIMENTAL SECTION
■
Reagents. All of the reagents except 5-bromo-8-amino-1,7-
naphthyridine (4) and 5-iodo-8-amino-1,7-naphthyridine (9)
were purchased and used without further purification.
Syntheses of 4 and 9 are described in the Supporting
Information.
NMR Analysis. Samples for 1H and 13C NMR spectroscopy
were analyzed using a Bruker AV-III-500-HD NMR
spectrometer.
High-Throughput Experiments. The HTE experiments
were performed using a Biomek i7 liquid handling robot
(Beckman Coulter, Inc., Indianapolis, IN). All of the methods
prepared utilized transfer of reagents dissolved in the desired
solvent from chemically resistant reservoirs to the heating
block. The span-8 pod in the robotic system, comprising 8
channels to deliver the desired volumes from source to
destination, was used. The pipetting techniques used in the
robot for transfer of volatile solvents were modified to dispense
accurate volumes. A multichannel pod on the robotic system
with a 384-tip head was used to transfer solutions from the 96-
well plate to the 384-well plate and for mixing prior to heating
H
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Microfluidic Synthesis. All of the microfluidic experi-
ments for screening of Sonogashira and amide coupling
reaction conditions were carried out using a Labtrix S1 system
(Chemtrix BV, Echt, The Netherlands) with staggered
oriented ridge 3227 reactor chips (19.5 μL). The system was
configured with five syringe pumps feeding a microreactor
positioned on a Peltier temperature control stage. FEP tubing
(0.8 mm o.d. × 0.25 mm i.d., Dolomite Microfluidics) with 1
mL gastight glass syringes (Hamilton, Reno, NV) were used.
The recipes for the reaction conditions were programmed
using the ChemTrix GUI software. For the scale-up of the
amidation and telescoped flow synthesis, PFA tubing (1/16″
o.d. × 0.03″ i.d., IDEX Health & Science) was coiled around
Cu elbow joints (Home Depot) and placed in the oil bath at
the desired temperature. All of the microfluidic parts, including
unions, superflangeless nuts, back-pressure regulators, and T-
mixers, were purchased from IDEX Health & Science.
the runs. Blind plugs were placed on the second and third
outlets. The final outlet was connected to the ultralow-volume
back-pressure regulator, followed by collection of the samples
into the autosampler. The flow rates for S1 and S2 were 0.594
and 0.186 μL/min, respectively, thereby producing a residence
time of 25 min. The flow rates for the pumps were changed to
achieve residence times of 15, 25, 30, 40, and 50 min according
to the recipe. For analysis of the yield or conversion using LC−
MS, 50 μL of the crude sample was collected, directly diluted
in ACN to 50 μM, and filtered using a 0.2 μm PTFE syringe
filter. For the isolated yield determination, solutions of the
crude product were extracted with 3× volume (i.e., 3 times the
volume of the crude solution collected) of DCM and washed
with 1× volumes of water five times. The combined organic
layers were washed with 1× volumes of saturated NH4Cl and
brine, dried over anhydrous Na2SO4, filtered, and concentrated
under reduced pressure. The crude mixture was purified by
column chromatography to obtain the desired product (Rf 0.2,
5-Ethynyl-N-(4-((4-methylpiperazin-1-yl)methyl)-3-
(trifluoromethyl)phenyl)nicotinamide (3). Amine 1
(151.63 mg, 0.55 mmol, 0.277 M), carboxylic acid 2 (80.92
mg, 0.55 mmol, 0.277 M), and DIPEA (287.4 μL, 1.65 mmol,
3 equiv) were combined to make a final solution in 2 mL of
DMF that was subsequently loaded into a 1 mL Hamilton
syringe mounted on pump 1 of the Chemtrix reactor. A 2 mL
solution of HATU (230.08 mg, 0.605 mmol, 1.1 equiv., 0.302
M) in DMF was prepared and loaded into another 1 mL
Hamilton syringe and mounted on the pump 2. The outlets of
pumps 1 and 2 were fed into a T-junction, from where the
mixture entered the reactor holder containing the 3227 reactor
chip (19.5 μL) placed on the Peltier stage. Inlets 2 and 3 were
blocked by blind plugs, and the outlet of the reactor was
channeled into the carousel unit. The outlet stream was
collected as waste for 3 times the residence time prior to
collection of the sample for every condition for which a recipe
was made on the software. Details of the residence times used
are provided in the Supporting Information. LC−MS was
performed directly on the crude reaction mixture by diluting it.
For TLC analysis, reaction mixtures were diluted 1:1 with
DCM. Product isolation for yield determination was achieved
by collecting 500 μL of crude reaction mixture, to which water
was added directly until a slurry was observed as the product
started to precipitate. The product was then gathered by
filtration and dried under vacuum overnight. TLC: Rf 0.5 (1:1
1
1:1 MeOH/acetone). The H and 13C NMR data matched
those reported in the literature.
Telescoped Continuous Flow Synthesis of 5. A 30 mL
stock solution in HPLC-grade DMF containing 4 (436.9 mg,
1.95 mmol, 0.065 M), CuI (3.61 mg, 0.019 mmol, 0.632 mM),
and pyrrolidine (457.57 μL, 5.57 mmol, 0.186 M) was
prepared using a volumetric flask, and the solution was then
loaded into a 25 mL Hamilton syringe (labeled A) using a 0.2
μm syringe filter. PdCl2(PPh3)2 (44 mg, 0.063 mmol, 2.09
mM) in 30 mL of DMF, HATU (2489.4 mg, 6.54 mmol,
0.2182 M) in 30 mL of DMF, and a 30 mL solution containing
carboxylic acid 2 (875.72 mg, 5.95 mmol, 0.1984 M), amine 1
(1626.68 mg, 5.952 mmol, 0.1984 M), and DIPEA (3110.36
μL, 17.85 mmol, 0.5952 M) in DMF were prepared, and these
solutions were loaded into Hamilton syringes labeled B, C, and
D, respectively, and mounted on the Harvard syringe pumps.
The flow rates for syringes A, B, C, D were set at 14.26, 4.326,
4.45, and 4.45 μL/min, respectively. Other flow rates are listed
in Table 3. Solutions of the crude product were extracted with
3× volume of DCM and washed with 1× volumes of water five
times. The combined organic layers were washed with 1×
volumes of saturated NH4Cl and brine, dried over anhydrous
Na2SO4, filtered, and concentrated under reduced pressure.
The crude mixture was purified by column chromatography to
obtain the desired product (Rf 0.2, 1:1 MeOH/acetone).
1
1
MeOH/acetone). The H NMR, 13C NMR, and HPLC data
HPLC and H and 13C NMR analyses of the product are
are provided in the Supporting Information.
provided in the Supporting Information.
Continuous Flow Synthesis of 5-((8-Amino-1,7-naph-
thyridin-5-yl)ethynyl)-N-(4-((4-methylpiperazin-1-yl)-
methyl)-3-(trifluoromethyl)phenyl)nicotinamide
(HSN608, 5). A 4 mL DMF solution of alkyne 3 (120.73 mg,
0.3 mmol, 0.075 M), compound 4 (70.57 mg, 0.315 mmol,
0.0787 M), CuI (0.57 mg, 0.003 mmol, 0.75 mM), and
pyrrolidine (75.12 μL, 0.915 mmol, 0.2287 M) was degassed
by passing Ar for 30 min and then loaded onto gastight glass
syringe port 1 (SI section 10) after filtering using a 0.2 μm
PTFE syringe filter. Syringe 2 was loaded with 1 mol % catalyst
(2.1 mg in the case of PdCl2(PPh3)2, 0.003 mmol) in DMF
after purging under Ar. [(t-Bu)3PH]BF4 was added at 3 mol %
when 3 mol % PdCl2 or PdCl2(MeCN)2 was used as the
catalyst in syringe 2). The solutions from the two syringe inlets
passed through check valves and were fed into a T-connector,
from which mixture flowed into the Chemtrix reactor block
mounted on the Peltier heating stage at the desired
temperature. The 3227 reactor chip (19.5 μL) was used for
Continuous Flow Synthesis of Amide Coupling for 3.
5-Ethynylnicotinic acid (1) (1.018 g, 6.925 mmol, 1 equiv),
amine 2 (1.892 g, 6.925 mmol, 1 equiv), and DIPEA (3.618
mL, 3 equiv) were added to a 25 mL volumetric flask that was
then filled with DMF. HATU (2.870 g) was added to a
separate 25 mL volumetric flask that was then filled with DMF.
After complete mixing of all of the reagents in the flasks, the
solutions were transferred to beakers and then loaded into
Hamilton syringes. The coiled reactor was placed on a hot
plate at 50 °C and equilibrated for 30 min. Prior to the
reaction, 10 mL of DMF was flowed through the PFA tubing.
The syringes with the respective solutions were mounted onto
the Harvard syringe pump (Ultra) where a flow rate of 17.8
μL/min was set. The reaction solution was collected for 3
times the residence time (i.e., 60 min for equilibration) before
switching to collect the crude reaction solution in a 50 mL
Falcon tube. Water was added to the crude reaction solution
until precipitation was observed. The precipitate was filtered
I
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
using a fritted funnel to obtain 2.260 g of the desired product
(83% yield).
discussions with Prof. R. Graham Cooks (Department of
Chemistry, Purdue University), Tiago J. P. Sobreira (CHRIS
software developer), and Larisa Avramova, Bruce Cooper,
Ahmed Mufti, and Ryan Hilger (Amy Facility at Purdue
University, Department of Chemistry, heating block fabrica-
tion).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00289.
■
sı
ABBREVIATIONS
Detailed experimental procedures for high-throughput
experimentation, ESI-MS, DESI-MS, batch synthesis,
continuous flow synthesis, telescoped synthesis, DoE
■
HTE, high-throughput experimentation; DoE, Design of
Experiments; DESI-MS, desorption electrospray ionization-
mass spectrometry; FLT-3, fms-like tyrosine kinase 3; OVAT,
one variable at a time; CHRIS, chemical reaction integrated
screening software; AML, acute myeloid leukemia; TLC, thin-
layer chromatography; LC−MS, liquid chromatography−mass
spectrometry; ESI-MS, electrospray ionization mass spectrom-
etry; NMR, nuclear magnetic resonance; CFI, coiled flow
inverter; RT, residence time; HATU, O-(7-azabenzotriazol-1-
yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate;
TMPH, 2,2,6,6-tetramethylpiperidine; TBAPF6, tetrabutylam-
monium hexafluorophosphate; Pd, palladium; Cu, copper;
DIPEA, diisopropylethylamine; TEA, triethylamine; TMEDA,
N,N,N′,N′-tetramethylethylenediamine; K3PO4, potassium
phosphate; ANOVA, analysis of variance; K2CO3, potassium
carbonate; PdCl2, palladium(II) chloride; [(t-Bu)3PH]BF4, tri-
tert-butylphosphonium tetrafluoroborate; PdCl2(PPh3)2, bis-
(triphenylphosphine)palladium(II) dichloride; Pd(amphos)-
Cl2, [bis(di-tert-butyl(4-dimethylaminophenyl)phosphine)-
dichloropalladium(II)]; XPhos Pd G3, (2-dicyclohexylphos-
phino-2′,4′,6′-triisopropyl-1,1′-biphenyl)[2-(2′-amino-1,1′-
biphenyl)]palladium(II) methanesulfonate; PdCl2(MeCN)2,
bis(acetonitrile)palladium dichloride; Pd(OAc)2, palladium-
(II) acetate; DMF, N,N′-dimethylformamide; EtOH, ethanol;
ACN, acetonitrile; DMAc, N,N′-dimethylacetamide; NMP, N-
methylpyrrolidone; DCM, dichloromethane; DMSO, dimethyl
sulfoxide; PFA, perfluoroalkoxy; SOR, staggered oriented ridge
1
analysis, H and 13C NMR spectra, and spectrometric
data (PDF)
AUTHOR INFORMATION
Corresponding Author
■
David H. Thompson − Department of Chemistry, Multi-
disciplinary Cancer Research Facility, and Institute for Drug
Discovery, Purdue University, West Lafayette, Indiana 47907,
United States; orcid.org/0000-0002-0746-1526;
Email: davethom@purdue.edu
Authors
Shruti A. Biyani − Department of Chemistry, Multi-disciplinary
Cancer Research Facility, and Institute for Drug Discovery,
Purdue University, West Lafayette, Indiana 47907, United
States; orcid.org/0000-0003-2756-3906
Qingqing Qi − Department of Chemistry, Multi-disciplinary
Cancer Research Facility, and Institute for Drug Discovery,
Purdue University, West Lafayette, Indiana 47907, United
States
Jingze Wu − Department of Chemistry, Multi-disciplinary
Cancer Research Facility, and Institute for Drug Discovery,
Purdue University, West Lafayette, Indiana 47907, United
States; orcid.org/0000-0002-7479-8284
Yuta Moriuchi − Department of Chemistry, Multi-disciplinary
Cancer Research Facility, and Institute for Drug Discovery,
Purdue University, West Lafayette, Indiana 47907, United
States
Elizabeth A. Larocque − Department of Chemistry, Multi-
disciplinary Cancer Research Facility, and Institute for Drug
Discovery, Purdue University, West Lafayette, Indiana 47907,
United States
Herman O. Sintim − Department of Chemistry, Multi-
disciplinary Cancer Research Facility, and Institute for Drug
Discovery, Purdue University, West Lafayette, Indiana 47907,
United States
REFERENCES
■
(1) (a) Krska, S. W.; DiRocco, D. A.; Dreher, S. D.; Shevlin, M. The
Evolution of Chemical High-Throughput Experimentation to Address
Challenging Problems in Pharmaceutical Synthesis. Acc. Chem. Res.
2017, 50, 2976−2985. (b) Lendrem, D.; Owen, M.; Godbert, S. DoE
(Design of Experiments) in Development Chemistry: Potential
Obstacles. Org. Process Res. Dev. 2001, 5, 324−327.
(2) (a) Milo, A. Democratizing Synthesis by Automation. Science
2019, 363, 122−123. (b) Tovar, M.; Mahler, L.; Buchheim, S.; Roth,
M.; Rosenbaum, M. A. Monitoring and External Control of pH in
Microfluidic Droplets during Microbial Culturing. Microb. Cell Fact.
2020, 19, 16. (c) Raveglia, L. F.; Giardina, G. A. Accelerating the
Drug-Discovery Process: New Tools and Technologies Available to
Medicinal Chemists. Future Med. Chem. 2009, 1, 1019−1023.
(d) Wiles, C.; Watts, P. Continuous-Flow Organic Synthesis: A
Tool for the Modern Medicinal Chemist. Future Med. Chem. 2009, 1,
1593−1612.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00289
Author Contributions
†S.A.B. and Q.Q. contributed equally to this work.
Notes
(3) (a) Larrosa-Garcia, M.; Baer, M. R. Flt3 Inhibitors in Acute
Myeloid Leukemia: Current Status and Future Directions. Mol. Cancer
Ther. 2017, 16, 991−1001. (b) Pratz, K. W.; Levis, M. How I Treat
Flt3-Mutated Aml. Blood 2017, 129, 565−571.
The authors declare the following competing financial
interest(s): H.O.S. is a cofounder of KinaRx LLC.
̈
(4) (a) Dohner, H.; Weisdorf, D. J.; Bloomfield, C. D. Acute
ACKNOWLEDGMENTS
■
Myeloid Leukemia. N. Engl. J. Med. 2015, 373, 1136−1152.
(b) Thein, M. S.; Ershler, W. B.; Jemal, A.; Yates, J. W.; Baer, M.
R. Outcome of Older Patients with Acute Myeloid Leukemia: An
Analysis of Seer Data over 3 Decades. Cancer 2013, 119, 2720−2727.
(c) Doehner, H.; Estey, E. H.; Amadori, S.; Appelbaum, F. R.;
Buechner, T.; Burnett, A. K.; Dombret, H.; Fenaux, P.; Grimwade, D.;
Larson, R. A.; Lo-Coco, F.; Naoe, T.; Niederwieser, D.;
This work was supported by the Defense Advanced Research
Projects Agency (Award W911NF-16-2-0020), the Center for
Cancer Research at Purdue University (P30 CA023168), and
Boilermaker Health Innovations. David L. Logsdon is
acknowledged for assistance with MRM method development
and DESI-MS. We appreciate the efforts and helpful
J
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Ossenkoppele, G. J.; Sanz, M. A.; Sierra, J.; Tallman, M. S.;
Loewenberg, B.; Bloomfield, C. D. Diagnosis and Management of
Acute Myeloid Leukemia in Adults: Recommendations from an
International Expert Panel, on Behalf of the European Leukemianet.
Blood 2010, 115, 453−474.
Experiments. J. Org. Chem. 2004, 69, 1097−1103. (d) Tye, H.
Application of Statistical ‘Design of Experiments’ Methods in Drug
Discovery. Drug Discovery Today 2004, 9, 485−491. (e) Denmark, S.
E.; Butler, C. R. Vinylation of Aromatic Halides Using Inexpensive
Organosilicon Reagents. Illustration of Design of Experiment
Protocols. J. Am. Chem. Soc. 2008, 130, 3690−3704. (f) Buitrago
Santanilla, A.; Regalado, E. L.; Pereira, T.; Shevlin, M.; Bateman, K.;
Campeau, L.-C.; Schneeweis, J.; Berritt, S.; Shi, Z.-C.; Nantermet, P.;
Liu, Y.; Helmy, R.; Welch, C. J.; Vachal, P.; Davies, I. W.; Cernak, T.;
Dreher, S. D. Nanomole-Scale High-Throughput Chemistry for the
Synthesis of Complex Molecules. Science 2015, 347, 49−53.
(5) Antar, A. I.; Otrock, Z. K.; Jabbour, E.; Mohty, M.; Bazarbachi,
A. Flt3 Inhibitors in Acute Myeloid Leukemia: Ten Frequently Asked
Questions. Leukemia 2020, 34, 682−696.
(6) (a) Naganna, N.; Opoku-Temeng, C.; Choi, E. Y.; Larocque, E.;
Chang, E. T.; Carter-Cooper, B. A.; Wang, M.; Torregrosa-Allen, S.
E.; Elzey, B. D.; Lapidus, R. G.; Sintim, H. O. Amino Alkynylisoquino-
line and Alkynylnaphthyridine Compounds Potently Inhibit Acute
Myeloid Leukemia Proliferation in Mice. EBioMedicine 2019, 40,
231−239. (b) De Kouchkovsky, I.; Abdul-Hay, M. Acute Myeloid
Leukemia: A Comprehensive Review and 2016 Update. Blood Cancer
J. 2016, 6, e441. (c) McCormick, S. R.; McCormick, M. J.; Grutkoski,
P. S.; Ducker, G. S.; Banerji, N.; Higgins, R. R.; Mendiola, J. R.;
Reinartz, J. J. Flt3 Mutations at Diagnosis and Relapse in Acute
Myeloid Leukemia: Cytogenetic and Pathologic Correlations,
Including Cuplike Blast Morphology. Arch. Pathol. Lab Med. 2010,
134, 1143−1151. (d) Gilliland, D. G.; Griffin, J. D. The Roles of Flt3
in Hematopoiesis and Leukemia. Blood 2002, 100, 1532−1542.
(7) Wang, M.; Naganna, N.; Sintim, H. O. Identification of
Nicotinamide Aminonaphthyridine Compounds as Potent Ret Kinase
Inhibitors and Antitumor Activities against Ret Rearranged Lung
Adenocarcinoma. Bioorg. Chem. 2019, 90, 103052.
(8) (a) Sperry, J. B.; Minteer, C. J.; Tao, J.; Johnson, R.; Duzguner,
R.; Hawksworth, M.; Oke, S.; Richardson, P. F.; Barnhart, R.; Bill, D.
R.; Giusto, R. A.; Weaver, J. D. Thermal Stability Assessment of
Peptide Coupling Reagents Commonly Used in Pharmaceutical
Manufacturing. Org. Process Res. Dev. 2018, 22, 1262−1275.
(b) Sneddon, H. F. Safety First. Green Chem. 2016, 18, 5082−5085.
(9) Larocque, E. A.; Naganna, N.; Opoku-Temeng, C.; Lambrecht,
A. M.; Sintim, H. O. Alkynylnicotinamide-Based Compounds as Abl1
Inhibitors with Potent Activities against Drug-Resistant Cml
Harboring Abl1(T315i) Mutant Kinase. ChemMedChem 2018, 13,
1172−1180.
(16) (a) Shevlin, M. Practical High-Throughput Experimentation for
Chemists. ACS Med. Chem. Lett. 2017, 8, 601−607. (b) Mennen, S.
M.; Alhambra, C.; Allen, C. L.; Barberis, M.; Berritt, S.; Brandt, T. A.;
́
Campbell, A. D.; Castanon, J.; Cherney, A. H.; Christensen, M.;
̃
Damon, D. B.; Eugenio de Diego, J.; García-Cerrada, S.; García-
Losada, P.; Haro, R.; Janey, J.; Leitch, D. C.; Li, L.; Liu, F.; Lobben, P.
C.; MacMillan, D. W. C.; Magano, J.; McInturff, E.; Monfette, S.;
Post, R. J.; Schultz, D.; Sitter, B. J.; Stevens, J. M.; Strambeanu, I. I.;
Twilton, J.; Wang, K.; Zajac, M. A. The Evolution of High-
Throughput Experimentation in Pharmaceutical Development and
Perspectives on the Future. Org. Process Res. Dev. 2019, 23, 1213−
1242. (c) Allen, C. L.; Leitch, D. C.; Anson, M. S.; Zajac, M. A. The
Power and Accessibility of High-Throughput Methods for Catalysis
Research. Nat. Catal. 2019, 2, 2−4. (d) Carson, N. Rise of the
Robots. Chem. - Eur. J. 2020, 26, 3194−3196. (e) Campos, K. R.;
Coleman, P. J.; Alvarez, J. C.; Dreher, S. D.; Garbaccio, R. M.; Terrett,
N. K.; Tillyer, R. D.; Truppo, M. D.; Parmee, E. R. The Importance of
Synthetic Chemistry in the Pharmaceutical Industry. Science 2019,
363, eaat0805.
(17) Wleklinski, M.; Loren, B. P.; Ferreira, C. R.; Jaman, Z.;
Avramova, L.; Sobreira, T. J. P.; Thompson, D. H.; Cooks, R. G. High
Throughput Reaction Screening Using Desorption Electrospray
Ionization Mass Spectrometry. Chem. Sci. 2018, 9, 1647.
̈
(18) Noel, T.; Buchwald, S. L. Cross-Coupling in Flow. Chem. Soc.
Rev. 2011, 40, 5010−5029.
(19) (a) Voltrova, S.; Srogl, J. Sonogashira Cross-Coupling under
Non-Basic Conditions. Flow Chemistry as a New Paradigm in
Reaction Control. Org. Chem. Front. 2014, 1, 1067−1071. (b) Tan,
L.-M.; Sem, Z.-Y.; Chong, W.-Y.; Liu, X.; Hendra; Kwan, W. L.; Lee,
C.-L. K. Continuous Flow Sonogashira C-C Coupling Using a
Heterogeneous Palladium-Copper Dual Reactor. Org. Lett. 2013, 15,
65−67. (c) Lange, P. P.; James, K. Rapid Access to Compound
Libraries through Flow Technology: Fully Automated Synthesis of a
3-Aminoindolizine Library Via Orthogonal Diversification. ACS
Comb. Sci. 2012, 14, 570−578. (d) Zhang, Y.; Jamison, T. F.; Patel,
S.; Mainolfi, N. Continuous Flow Coupling and Decarboxylation
Reactions Promoted by Copper Tubing. Org. Lett. 2011, 13, 280−
283. (e) Placzek, M. S.; Chmielecki, J. M.; Houghton, C.; Calder, A.;
Wiles, C.; Jones, G. B. Fluoroalkynylations of Aryl Halides under
Continuous-Flow Homogeneous Catalysis. J. Flow Chem. 2013, 3,
46−50. (f) Battilocchio, C.; Bhawal, B. N.; Chorghade, R.; Deadman,
B. J.; Hawkins, J. M.; Ley, S. V. Flow-Based, Cerium Oxide Enhanced,
Low-Level Palladium Sonogashira and Heck Coupling Reactions by
Perovskite Catalysts. Isr. J. Chem. 2014, 54, 371−380. (g) Teci, M.;
Tilley, M.; McGuire, M. A.; Organ, M. G. Using Anilines as Masked
Cross-Coupling Partners: Design of a Telescoped Three-Step Flow
Diazotization, Iododediazotization, Cross-Coupling Process. Chem. -
Eur. J. 2016, 22, 17407−17415. (h) Znidar, D.; Hone, C. A.; Inglesby,
P.; Boyd, A.; Kappe, C. O. Development of a Continuous-Flow
Sonogashira Cross-Coupling Protocol Using Propyne Gas under
Process Intensified Conditions. Org. Process Res. Dev. 2017, 21, 878−
884. (i) Sugimoto, A.; Fukuyama, T.; Rahman, M. T.; Ryu, I. An
Automated-Flow Microreactor System for Quick Optimization and
Production: Application of 10- and 100-Gram Order Productions of a
Matrix Metalloproteinase Inhibitor Using a Sonogashira Coupling
Reaction. Tetrahedron Lett. 2009, 50, 6364−6367. (j) Kawanami, H.;
Matsushima, K.; Sato, M.; Ikushima, Y. Rapid and Highly Selective
Copper-Free Sonogashira Coupling in High-Pressure, High-Temper-
(10) Gutmann, B.; Cantillo, D.; Kappe, C. O. Continuous-Flow
Technologya Tool for the Safe Manufacturing of Active
Pharmaceutical Ingredients. Angew. Chem., Int. Ed. 2015, 54, 6688−
6728.
(11) Britton, J.; Raston, C. L. Multi-Step Continuous-Flow
Synthesis. Chem. Soc. Rev. 2017, 46, 1250−1271.
(12) Cole, K. P.; Johnson, M. D. Continuous Flow Technology Vs.
The Batch-by-Batch Approach to Produce Pharmaceutical Com-
pounds. Expert Rev. Clin. Pharmacol. 2018, 11, 5−13.
(13) Garrett, C. E.; Prasad, K. The Art of Meeting Palladium
Specifications in Active Pharmaceutical Ingredients Produced by Pd-
Catalyzed Reactions. Adv. Synth. Catal. 2004, 346, 889−900.
́
́
(14) (a) Jimenez-Gonzalez, C.; Poechlauer, P.; Broxterman, Q. B.;
Yang, B.-S.; am Ende, D.; Baird, J.; Bertsch, C.; Hannah, R. E.;
Dell’Orco, P.; Noorman, H.; Yee, S.; Reintjens, R.; Wells, A.;
Massonneau, V.; Manley, J. Key Green Engineering Research Areas
for Sustainable Manufacturing: A Perspective from Pharmaceutical
and Fine Chemicals Manufacturers. Org. Process Res. Dev. 2011, 15,
900−911. (b) Hartman, R. L.; McMullen, J. P.; Jensen, K. F. Deciding
Whether to Go with the Flow: Evaluating the Merits of Flow Reactors
for Synthesis. Angew. Chem., Int. Ed. 2011, 50, 7502−7519.
̌
(15) (a) Murray, P. M.; Bellany, F.; Benhamou, L.; Bucar, D.-K.;
Tabor, A. B.; Sheppard, T. D. The Application of Design of
Experiments (Doe) Reaction Optimisation and Solvent Selection in
the Development of New Synthetic Chemistry. Org. Biomol. Chem.
2016, 14, 2373−2384. (b) Owen, M. R.; Luscombe, C.; Lai; Godbert,
S.; Crookes, D. L.; Emiabata-Smith, D. Efficiency by Design:
Optimisation in Process Research. Org. Process Res. Dev. 2001, 5,
308−323. (c) Stazi, F.; Palmisano, G.; Turconi, M.; Clini, S.;
Santagostino, M. Accelerated Koenigs−Knorr Glucuronidation of a
Deactivated Nitrophenol: Unveiling the Role of Polyamine Additive
1,1,4,7,10,10-Hexamethyltriethylenetetramine1 through Design of
K
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
ature Water in a Microfluidic System. Angew. Chem., Int. Ed. 2007, 46,
5129−5132.
(20) (a) Spare, L. K.; Laude, V.; Harman, D. G.; Aldrich-Wright, J.
R.; Gordon, C. P. An Optimised Approach for Continuous-Flow
Solid-Phase Peptide Synthesis Utilising a Rudimentary Flow Reactor.
React. Chem. Eng. 2018, 3, 875−882. (b) Bogdan, A. R.; Wang, Y. A
High-Throughput Synthesis of 1,2,4-Oxadiazole and 1,2,4-Triazole
Libraries in a Continuous Flow Reactor. RSC Adv. 2015, 5, 79264−
79269. (c) Mijalis, A. J.; Thomas, D. A., III; Simon, M. D.; Adamo, A.;
Beaumont, R.; Jensen, K. F.; Pentelute, B. L. A Fully Automated Flow-
Based Approach for Accelerated Peptide Synthesis. Nat. Chem. Biol.
2017, 13, 464−466. (d) Gordon, C. P. The Renascence of
Continuous-Flow Peptide Synthesis − an Abridged Account of
Solid and Solution-Based Approaches. Org. Biomol. Chem. 2018, 16,
180−196.
(21) (a) Byrne, F. P.; Jin, S.; Paggiola, G.; Petchey, T. H. M.; Clark,
J. H.; Farmer, T. J.; Hunt, A. J.; Robert McElroy, C.; Sherwood, J.
Tools and Techniques for Solvent Selection: Green Solvent Selection
Guides. Sustainable Chem. Processes 2016, 4, 7. (b) Prat, D.; Hayler, J.;
Wells, A. A Survey of Solvent Selection Guides. Green Chem. 2014,
16, 4546−4551.
(22) Albericio, F.; Bofill, J. M.; El-Faham, A.; Kates, S. A. Use of
Onium Salt-Based Coupling Reagents in Peptide Synthesis1. J. Org.
Chem. 1998, 63, 9678−9683.
(23) (a) Jaman, Z.; Mufti, A.; Sah, S.; Avramova, L.; Thompson, D.
H. High Throughput Experimentation and Continuous Flow
Validation of Suzuki−Miyaura Cross-Coupling Reactions. Chem. -
Eur. J. 2018, 24, 9546−9554. (b) de Vries, J. G.; de Vries, A. H. M.
The Power of High-Throughput Experimentation in Homogeneous
Catalysis Research for Fine Chemicals. Eur. J. Org. Chem. 2003, 2003,
799−811.
(24) Gallagher, P. K.; Gross, M. E. The Thermal Decomposition of
Palladium Acetate. J. Therm. Anal. 1986, 31, 1231−1241.
(25) (a) Bruno, N. C.; Buchwald, S. L. Synthesis and Application of
Palladium Precatalysts That Accommodate Extremely Bulky Di-Tert-
Butylphosphino Biaryl Ligands. Org. Lett. 2013, 15, 2876−2879.
(b) Surry, D. S.; Buchwald, S. L. Dialkylbiaryl Phosphines in Pd-
Catalyzed Amination: A User’s Guide. Chem. Sci. 2011, 2, 27−50.
(26) (a) Labinger, J. A. Tutorial on Oxidative Addition. Organo-
́
metallics 2015, 34, 4784−4795. (b) Chinchilla, R.; Najera, C. The
Sonogashira Reaction: A Booming Methodology in Synthetic Organic
Chemistry. Chem. Rev. 2007, 107, 874−922.
̌
(27) (a) Gazvoda, M.; Virant, M.; Pinter, B.; Kosmrlj, J. Mechanism
of Copper-Free Sonogashira Reaction Operates through Palladium-
Palladium Transmetallation. Nat. Commun. 2018, 9, 4814. (b) Chin-
́
chilla, R.; Najera, C. Recent Advances in Sonogashira Reactions.
Chem. Soc. Rev. 2011, 40, 5084−5121. (c) Gelman, D.; Buchwald, S.
L. Efficient Palladium-Catalyzed Coupling of Aryl Chlorides and
Tosylates with Terminal Alkynes: Use of a Copper Cocatalyst Inhibits
the Reaction. Angew. Chem., Int. Ed. 2003, 42, 5993−5996.
(28) (a) He, C.; Ke, J.; Xu, H.; Lei, A. Synergistic Catalysis in the
Sonogashira Coupling Reaction: Quantitative Kinetic Investigation of
Transmetalation. Angew. Chem., Int. Ed. 2013, 52, 1527−1530.
(b) Wang, X.; Song, Y.; Qu, J.; Luo, Y. Mechanistic Insights into the
Copper-Cocatalyzed Sonogashira Cross-Coupling Reaction: Key Role
of an Anion. Organometallics 2017, 36, 1042−1048.
(29) Soni, S.; Sharma, L.; Meena, P.; Roy, S.; Nigam, K. D. P.
Compact Coiled Flow Inverter for Process Intensification. Chem. Eng.
Sci. 2019, 193, 312−324.
(30) Tan, F. H.; Putoczki, T. L.; Stylli, S. S.; Luwor, R. B. Ponatinib:
A Novel Multi-Tyrosine Kinase Inhibitor against Human Malig-
nancies. OncoTargets Ther. 2019, 12, 635−645.
L
https://dx.doi.org/10.1021/acs.oprd.0c00289
Org. Process Res. Dev. XXXX, XXX, XXX−XXX