Identification of 2,2-Dimethylbutanoic Acid (HST5040), a Clinical Development Candidate for the Treatment of Propionic Acidemia and Methylmalonic Acidemia

and Methylmalonic Acidemia

Article

Identiﬁcation of 2,2-Dimethylbutanoic Acid (HST5040), a Clinical

Development Candidate for the Treatment of Propionic Acidemia

Allison J. Armstrong, Brad R. Henke, Maria Sol Collado, Justin M. Taylor, Taylor D. Pourtaheri,

John E. Dillberger, Thomas D. Roper, Brian R. Wamho ﬀ , Matthew W. Olson, Robert A. Figler,

Stephen A. Hoang, John E. Reardon, and Brian A. Johns*

Cite This: J. Med. Chem. 2021, 64, 5037 −5048

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: Propionic acidemia (PA) and methylmalonic acidemia (MMA) are rare

autosomal recessive disorders of propionyl-CoA (P-CoA) catabolism, caused by a

deﬁciency in the enzymes P-CoA carboxylase and methylmalonyl-CoA (M-CoA) mutase,

respectively. PA and MMA are classiﬁed as intoxication-type inborn errors of metabolism

because the intramitochondrial accumulation of P-CoA, M-CoA, and other metabolites

results in secondary inhibition of multiple pathways of intermediary metabolism, leading to

organ dysfunction and failure. Herein, we describe the structure−activity relationships of a

series of short-chain carboxylic acids which reduce disease-related metabolites in PA and

MMA primary hepatocyte disease models. These studies culminated in the identiﬁcation of

,2-dimethylbutanoic acid (10, HST5040) as a clinical candidate for the treatment of PA

and MMA. Additionally, we describe the in vitro and in vivo absorption, distribution,

metabolism, and excretion proﬁle of HST5040, data from preclinical studies, and the synthesis of the sodium salt of HST5040 for

clinical trials.

INTRODUCTION

PA is caused by a deﬁciency in the enzyme propionyl-CoA

carboxylase (PCC), while MMA is caused by a deﬁciency in

the enzyme methylmalonyl-CoA mutase (MUT) or in

enzymes involved in the synthesis of its cofactor, 5′-

■

Propionic acidemia (PA) and methylmalonic acidemia

MMA) are extremely rare, autosomal recessive intoxication-

type disorders aﬀecting sequential steps in the same propionate

(

deoxyadenosyl-cobalamin (vitamin B12). The reactions

catabolic pathway, leading to similar acute and chronic disease

catalyzed by PCC and MUT represent critical steps in

propionic acid catabolism, funneling metabolites from the

breakdown of the branched-chain amino acids valine and

isoleucine, as well as methionine, threonine, odd-chain fatty

acids, and the side chain of cholesterol, into the tricarboxylic

acid (TCA) cycle intermediate, succinyl-CoA (Figure 1).

The pathophysiology of PA and MMA is characterized by

the intramitochondrial accumulation of propionyl-CoA (P-

CoA) and methylmalonyl-CoA (M-CoA), respectively, and

accumulation of derived organic acids such as 3-hydroxypro-

pionic acid, 2-methyl citric acid (MCA), and methylmalonic

acid. These organic acids and other metabolites of P-CoA,

including propionyl carnitine (C3) and propionyl-glycine in

PA and methylmalonyl-carnitine in MMA, are present in

manifestations. PA and MMA are multisystemic diseases

aﬀecting central nervous system (CNS), renal, gastrointestinal

GI), immune, hepatic, hematologic, and cardiovascular

(

systems. Both diseases are associated with signiﬁcant morbidity

and mortality in infancy and childhood and for survivors,

signiﬁcant debilitating end-organ damage and neurological

issues in adulthood. Most PA and MMA patients present in the

neonatal period with an acute metabolic decompensation. This

decompensation is characterized by metabolic acidosis and

hyperammonemia that leads to lethargy, poor feeding, and

vomiting, which can progress to encephalopathy, coma, and

death if not identiﬁed and managed urgently. Many PA and

MMA patients experience frequent additional acute metabolic

decompensations associated with acute illness, fasting, or other

stressors such as surgical or interventional procedures even

with optimal disease management, which includes dietary

protein restriction, supplementation with nonpropiogenic

amino acids, carnitine supplementation, and the use of drugs

to lower ammonia levels. Each acute metabolic decompensa-

tion event likely contributes to accumulated chronic

Received: January 22, 2021

Published: April 13, 2021

complications over the course of life.

2021 The Authors. Published by

American Chemical Society

Journal of Medicinal Chemistry

037

J. Med. Chem. 2021, 64, 5037−5048

Article

screening allowed for the contribution of multiple acyl-CoA

synthetase family members which have selectivity toward

carboxylic acids of diﬀerent acyl chain lengths and steric bulk.

Additionally, the use of a phenotypic disease model for

screening allowed for identiﬁcation of compounds that could

potentially impact either P-CoA formation or clearance

pathways (or both). Table 1 summarizes the results of this

study.

Initially, compounds were tested at a single high

concentration (100 μM) for their ability to lower P-CoA in

this assay. Of the 30 compounds tested, 20 showed the ability

to lower P-CoA by at least 40% at the 100 μM concentration.

Several structure−activity trends emerged from this single-

concentration proﬁling work. First, compounds containing a

phenyl ring were uniformly ineﬀective at lowering P-CoA at

Figure 1. Biochemistry of PA and MMA disorders.

00 μM concentration (compounds 15−17, Table 1). Second,

plasma, urine, and other body ﬂuids. Therefore, a therapy

targeting the primary metabolic derangementthe accumu-

lation of P-CoA and M-CoAis expected to result in

reduction in the downstream metabolites implicated in the

toxicities which lead to clinical disease manifestations.

apart from 5,5,5-triﬂuoropentanoic acid (compound 20), all

compounds containing ﬂuorine atoms were inactive (com-

pounds 18, 19, and 21). Third, both mono- and di-alkyl

substitutions were tolerated at both the α- and β-positions of

the carboxylic acid. Finally, steric bulk at the γ-position is not

tolerated, as 4,4-dimethylpentanoic acid (compound 7) is

inactive.

As part of our ongoing eﬀorts to identify therapies for PA

and MMA, we hypothesized that low-molecular-weight

carboxylic acids that can readily form CoA esters using native

Compounds that were able to lower P-CoA by at least 40%

were then proﬁled in a full concentration−response manner to

determine both potency and eﬃcacy of the response (Table 1).

human cellular enzymes could lead to a redistribution of the

cellular acyl-CoA pools, thus potentially reducing abnormally

high acyl-CoAs such as P-CoA and M-CoA found in PA and

MMA patients. This report details the structure−activity

relationship (SAR) of a series of low-molecular-weight

carboxylic acids with respect to their ability to lower P-CoA

levels in primary hepatocytes derived from livers explanted

from PA patients. Based on these data and other factors

outlined below, we selected 2,2-dimethylbutanoic acid

Compound potency, quantiﬁed by IC₅₀or EC , is commonly

used to compare the performance of compounds in

pharmacological studies. However, it has been demonstrated

that these estimates are sensitive to technical noise and

deviations from biological and experimental assumptions. For

example, failure to experimentally observe the asymptotic

portion of a dose response can produce high uncertainty in the

parameter estimates of the logistic regression model from

which an IC₅₀or EC₅₀estimate is derived. Results from

cytotoxicity studies suggest that metrics based on the area

under the dose−response curve (AUDRC) may be more

reproducible indicators of compound activity than the

(

compound 10 in Table 1, referred to herein as HST5040)

for additional pharmacological proﬁling and ultimately as a

clinical development candidate for the treatment of PA and

MMA. This report also describes the detailed in vitro and in

vivo absorption, distribution, metabolism, and excretion

10,11

(

ADME) proﬁle of HST5040, additional data from preclinical

EC .

Furthermore, AUDRC and related values capture

studies that support selection of HST5040 as a clinical

candidate, and the synthesis of the sodium salt of HST5040,

termed HST5040A.

information about compound eﬃcacy in addition to potency,

which may be viewed as an advantage over the EC₅₀in some

applications. Thus, in this study, we employ a modiﬁcation of

the AUDRC, which we term the “adjusted AUDRC.” The

modiﬁcation involves the introduction of a penalty based on

the amount of variability in the data. This penalty decreases the

likelihood of favorable scores that spuriously arise from curves

ﬁt to noisy data (i.e., false positives). For our compounds,

adjusted AUDRC values near zero indicate compounds that

are highly eﬃcacious and potent in lowering P-CoA, while

values near one indicate inactive compounds. Due to

experimental variability in the PA patient-derived pHeps

used in this study, we could not calculate some compound

EC₅₀values with a high degree of robustness in repeat tests.

When the robustness of the data allowed for conﬁdence in

calculating an accurate EC value for a compound, those were

RESULTS AND DISCUSSION

■

The P-CoA-lowering activity of a set of commercially available

low-molecular-weight carboxylic acids was studied in a cell-

based phenotypic screen using a disease model of PA that

accumulates the primary metabolite of interest, P-CoA. This

model utilizes primary hepatocytes (pHeps) derived from

livers explanted from PA donors at the time of liver transplant

(

CNHS IRB Pro0004911). The PA pHeps have been

extensively characterized, including genotyping, assessment of

PCC protein expression, and the quantiﬁcation of intracellular

metabolite levels and other relevant disease biomarkers. For

this phenotypic screening assay, PA patient-derived pHeps

were treated for 1 h with individual compounds in the presence

calculated and are included in Table 1 (see Section S2 for

more details adjusted AUDRC calculations). Representative

concentration−response curves for all active compounds in

Table 1 and a graphical summary of normalized adjusted

AUDRC values for all compounds are found in the Supporting

Information (Figures S1 and S3 ).

of C-isoleucine, which is catabolized to form C-P-CoA. Of

note, compound eﬃcacy was similar when other sources of P-

CoA were utilized (data not shown). The reduction in C-P-

CoA and formation of each compound-CoA ester were

measured by MS/MS. We hypothesized that the formation

of a compound-CoA ester was important for the mechanism of

action, and therefore, the use of a cell-based model for

Gratifyingly, the compound activity seen at the 100 μM

concentration reproduced well in the concentration−response

038

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

Table 1. Structure−Activity Relationships of Compounds 1−30

039

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

Table 1. continued

Values are from at least n = 2 replicates (mean ± st. dev.) except where indicated. Only n = 1 replicate had a calculable value; IA = inactive,

deﬁned as <40% P-CoA lowering; ND = not done.

experiments. Additional SAR information can be derived from

examination of the adjusted AUDRC values from these

experiments. Branching at the α-position is not only tolerated

but it also generally leads to more potent and eﬃcacious

compounds. In fact, substitution with geminal dimethyl groups

in the α- or β-positions led to compounds with three of the

four lowest adjusted AUDRC values (compounds 6, 10, and

28) and the amide analogue 2,2-dimethylbutyramide (com-

pound 29). These three compounds cannot undergo acyl-CoA

formation unless ﬁrst biotransformed into their corresponding

carboxylic acids, so we were curious to see how they would

behave in this phenotypic assay. All three compounds appeared

active in the initial screening assay, but compounds 29 and 30

were inactive upon retesting in a concentration−response

fashion. However, methyl 2,2-dimethylbutanoate (28) proved

to be quite active, suggesting that this compound is readily

converted to the carboxylic acid by liver esterases in these

primary hepatocytes. It is not certain why this process appears

more facile than that for compound 30, but presumably, it is

due to their diﬀerences as substrates for hepatocyte esterases.

Importantly, as mentioned above, we also measured the

formation of the corresponding acyl-CoA for each of the test

compounds in Table 1. Example concentration−response

curves for both P-CoA reduction and the corresponding

acyl-CoA formation of compounds 2, 6, 10, and 11 can be

found in Section S4. In general, the potency of inhibition of P-

CoA showed good correlation with the potency of formation of

the corresponding compound acyl-CoA. Compounds that did

not lower P-CoA in this assay also did not have signiﬁcantly

elevated levels of their corresponding CoA esters, suggesting

either little formation of the CoA ester or rapid metabolism by

11). Compounds 22−25 containing heteroatoms in the side

chain were all classiﬁed as active in the single-shot assay;

however, in these concentration−response studies, they

showed generally high adjusted AUDRC values, indicating

poor potency relative to compounds with all-carbon side

chains. The lone exception was compound 23 which contains a

geminal dimethyl group in the α-position, reinforcing the

observation that branching at this position improves activity.

Carbocyclic analogues 8 and 12−14, although active, had

higher adjusted AUDRC values than acyclic analogues. In

summary, compounds that showed the most robust ability to

lower P-CoA in this assay have branching at the α- or β-

position and at least three carbons in the longest linear chain,

with compounds 2, 6, 10, 11, and 23 being the most

eﬃcacious.

We also proﬁled the methyl esters of valproic acid

compound 30) and 2,2-dimethylbutyric acid (compound

(

040

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

β-oxidation pathways. This implies a mechanistic link between

compound acyl-CoA formation and stability and P-CoA

reduction. Furthermore, we looked at the ammonia scavengers,

phenylbutyrate (17) and phenylacetate (15), which are used to

treat hyperammonemia by nitrogen scavenging and excretion

through CoA-dependent enzymatic reactions, ultimately

Table 2. HST5040 Biomarker Data in the HemoShear

Hepatocyte Bioreactor

disease

pHeps

% change at

biomarker

EC₅₀(μM)

30 μM

propionyl-CoA (P-CoA)

MMA

1.9 ± 1

2.1 ± 1.5

−78.8 ± 10.9

−74.2 ± 11.6

12,13

forming a phenacyl−glutamine conjugate.

We found no

methylmalonyl-CoA

reduction in P-CoA with either of these compounds,

suggesting the glutamine nitrogen scavenging pathway to be

unrelated to our proposed mechanism.

(

M-CoA)

MMA

normal

MMA

normal

2 ± 1.2

2.6 ± 1.9

1.4 ± 0.5

4.8 ± 0.2

4.3 ± 2.2

1.1 ± 0.5

0.8 ± 0.5

2.3 ± 1.9

−55 ± 6.6

−69.4 ± 16.2

−65.9 ± 10.7

54 ± 80.6

44.6 ± 26.8

−78.6 ± 12.9

−66.7 ± 14.9

54 ± 80.6

13.5 ± 24.9

20 ± 2.9

−31.3 ± 24.2

−22.3 ± 22.3

−35.2 ± 3

propionyl carnitine (C3)

acetyl carnitine (C2)

2-Methylcitric acid (MCA)

acetyl-CoA

From these results, 2,2-dimethylbutanoic acid (10,

HST5040) was selected for more thorough evaluation based

on three factors. First, this compound was among the most

potent and eﬃcacious at reducing the level of P-CoA in the

pHeps derived from PA patients from the list of compounds

we proﬁled, with an EC = 0.94 μM and an adjusted AUDRC

0.38. Second, the presence of the quaternary α-carbon of 2,2-

dimethylbutanoic acid prevents metabolism via the canonical

β-oxidation pathway. This avoids the formation of a reactive

and potentially toxic α,β-unsaturated carboxylate species that

has been observed to form with other small, saturated organic

acids that have an α-proton, such as 2-propylpentanoic

CoASH

11.6 ± 1.2

14.1 ± 8.4

3 ± 1.5

3.5 ± 1.7

8.3 ± 8

HST5040-CoA

(

valproic) acid. Third, during the course of our pharmacol-

MMA

normal

ogy studies, we became aware that human clinical safety

information on 2,2-dimethylbutanoic acid was available in the

public domain from previous clinical evaluation of this

compound which was under investigation for treatment of

Values are from an average of at least three replicates ± standard

deviation; NA = not applicable; NC = value could not be calculated.

15−19

sickle cell disease and β-thalassemia.

This previous work

(

CoASH) in these experiments, in general, there is a modest

was based on a mechanism of action unrelated to our current

pharmacology studies and disease indications; however, the

safety and tolerability data of 2,2-dimethylbutanoic acid

described in these phase I and phase II clinical trials were

favorable at doses that were projected to be above those

needed for treatment of PA and MMA patients.

eﬀect on CoASH and acetyl-CoA, suggesting that CoASH

levels and overall CoA adduct pools (as exempliﬁed by acetyl-

CoA) will not be substantially altered during chronic HST5040

therapy. The latter data provide a measure of reassurance that

these important CoA pools can be maintained in a range that

supports physiological homeostasis. These key disease

biomarkers are signiﬁcantly elevated (e.g. approximately 20-

fold for P-CoA and 40−80-fold for C3) in pHeps from patients

We have previously developed and utilized a proprietary in

vitro physiological model of human disease by applying

principles of physiological blood ﬂow to tissue from patients

with PA and MMA relative to pHeps from normal livers.

in order to better recapitulate their disease. This technology

has been used successfully to interrogate drug candidates and

HST5040 treatment results in a reduction of these biomarkers

to levels approaching normal values in this system, with P-CoA

within threefold and C3 within ninefold of baseline values seen

21−23

marketed drugs in multiple cellular systems,

and we have

recently applied it to recapitulate PA and MMA disease biology

in nondisease hepatocytes.

in primary hepatocytes derived from PA and MMA patients.

The pharmacology demonstrated by HST5040 in the PA

and MMA disease models was suﬃciently robust and

encouraging to enable further preclinical development of the

compound. We sought to determine the activity of HST5040

in pHeps from several potential species for nonclinical studies.

Since the reduction in P-CoA appears to be linked to the

biotransformation of HST5040 to HST5040-CoA, we used the

formation of HST5040-CoA as a measure of HST5040 activity.

The EC₅₀values for the formation of HST5040-CoA were

determined in pHeps from humans, mice, rats, beagle dogs,

minipigs, and cynomolgus monkeys and are presented in Table

3. The data revealed signiﬁcant interspecies diﬀerences in

potency of the formation of HST5040-CoA. The implications

were twofold. First, the substantially diminished potency in

mouse hepatocytes precluded the use of available PA and

We recently published a full characterization of the

pharmacology of HST5040 in this system, along with a

detailed discussion of the proposed mechanism of action of

HST5040 that explains the experimental observations.

Brieﬂy, PA, MMA, and normal human primary hepatocytes

were exposed to HST5040 (concentrations ranging from 0.1 to

00 μM) over the course of 6 days and key disease biomarkers

were measured at the completion of the experimental runs. A

summary of the key biomarker data is shown in Table 2 (for

concentration−response curves, see Supporting Information,

Figure S5 ).

The salient conclusions that can be drawn from these data

are as follows: (1) treatment of PA and MMA pHeps with

HST5040 under the more physiologically relevant “hepatocyte

bioreactor” culture conditions resulted in a signiﬁcant and

dose-dependent reduction in key PA and MMA disease

biomarkers P-CoA, M-CoA, MCA, and C3/C2 ratio, with

similar EC₅₀values for each biomarker, (2) HST5040 is

biotransformed into HST5040-CoA at an EC₅₀similar to that

for the reduction of P-CoA, and (3) although there is

signiﬁcant variability in both acetyl CoA and free coenzyme A

26,27

MMA mouse models

as an additional in vivo preclinical

eﬃcacy readout due to an inability to achieve and sustain

plasma levels above the EC₅₀in mice. Second, the data

indicated that the most relevant species for nonclinical studies

with HST5040 was the minipig.

Additional in vitro ADME proﬁling was conducted on

HST5040. Calculated and measured physicochemical proper-

041

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

Table 3. Interspecies Comparison of the Formation of

HST5040-CoA in Hepatocytes

ADME data reported below were obtained using HST5040A

as the test article.

The extent of protein binding of HST5040 in CD rat,

species

EC₅₀(μM)

Gottingen minipig, and human plasma was determined via

human

mouse

rat

5.9 ± 1.0

equilibrium dialysis. [ C]-HST5040A, at 40 μM concen-

130 ± 31.5

tration, was incubated with rat, minipig, and human plasma at

202

[

7 °C for 4 h in a dialysis device against phosphate buﬀer.

beagle dog

cynomolgus monkey

minipig

C]-HST5040 exhibited no protein binding in rat or human

46.2 ± 3.4

8.3 ± 2.3

plasma and low protein binding in minipig plasma (45%).

Additionally, the partitioning of HST5040 in CD rat,

Values are from n = 3 experiments unless otherwise indicated. n =

determined. [ C]-HST5040 exhibited low partitioning into

erythrocytes of CD rat, Gottingen minipig, and human blood

KB/P ≤ 1). There was >1-fold diﬀerence of the ratio of test

ttingen minipig, and human blood versus plasma was

ties of HST5040 are shown in Table 4. HST5040 was

evaluated as both a substrate and an inhibitor in Caco-2 and

(

compound concentrations in erythrocytes over plasma (KE/P)

between the concentrations in minipig blood observed;

however, the KE/P was ≤1, indicating low partitioning into

erythrocytes.

Table 4. Properties of HST5040

The in vivo pharmacokinetic proﬁles of HST5040 in mice,

rats, and minipigs given doses of HST5040A are summarized

in Table 5. HST5040 PK proﬁles after single doses of

HST5040A by the oral (po) and intravenous (iv) route were

assessed in CD-1 mice and Sprague−Dawley (CD) rats.

HST5040 generally exhibited similar PK properties in mice

and rats. Plasma clearance rates were low to moderate in both

species (ca. 8 and 26% of hepatic blood ﬂow, respectively).

Oral bioavailability (% F) was moderate to high in both species

and appeared to increase with dose; however, this was likely

attributable to the dose-dependent decrease in clearance. The

time to maximum plasma concentration (T_m_a_x) following oral

administration was generally short (0.25−0.5 h), indicating

rapid absorption. Upon escalation of the oral dose level,

maximum plasma concentration (Cmax) and area under the

curve (AUC) generally increased proportionally or slightly

more than proportionally. The elimination half-life (t1/2) was

<2 h in mice and approximately 3 h in rats.

MW/cLogP/PSA

pK_a

116.16/1.59/37.5

4.4

>20 × 10⁻cm/s, (<1)

>20 × 10⁻cm/s

>50 mg/mL@pH 7

0, 45, 0

permeability (Caco-2) (ER)

permeability (MCDKII-MDR1)

solubility

PPB (rat, minipig, human)

blood/plasma ratio (rat, minipig, human)

erythrocyte/plasma ratio (rat, minipig, human) 0.649, 0.195, 0.343

0.846, 0.605, 0.737

Data are for HST5040A. Data are for [ C]-HST5040A.

Madin−Darby canine kidney II cell line (MDCKII-MDR1)

cell monolayer assays. HST5040 is a high permeability

compound in both cell lines with apparent permeability

−

coeﬃcient (P ) values >20 × 10 in both apical to basal and

app

basal to apical directions. HST5040 is likely not a transported

substrate in either cell line as the eﬄux ratios were below 1.

HST5040 itself is poorly soluble in aqueous media; however,

the intended clinical dosage form of HST5040, which is the

sodium salt of HST5040 (termed HST5040A), displays

excellent solubility in pH 7-buﬀered water. Most of the

To assess tissue distribution and excretion of HST5040, a

quantitative whole-body autoradiography (QWBA) study

using radiolabeled HST5040A was performed in rats after a

single PO dose of [ C]-HST5040A at 10 mg/kg (100 μCi/

Table 5. In Vivo Plasma Pharmacokinetic Summary of HST5040A

dose

(mg/kg)

species

mouse

route

C_m_a_x(ng/mL)

T_m_a_x(h)

t₁_/₂(h)

V_d(L/kg)

(mL/min/kg)

AUC₀_‑_i_n_f(ng h/mL)

F (%)

59,425 ± 8552

33,725 ± 4656

11,8250 ± 9535

464,750 ± 95,963

---

0.125 ± 0.084

0.438 ± 0.125

0.5 ± 0.0

---

0.831 ± 0.376

0.59 ± 0.16

1.09 ± 0.931

1.91 ± 0.521

3.93 ± 1.26

3.17 ± 0.238

3.27 ± 0.138

2.73 ± 0.553

4.09 ± 1.03

---

0.469 ± 0.074

---

7.28 ± 1.27

70,155 ± 12,088

37,512 ± 7397

218,723 ± 69,960

1,245,621 ± 199,227

9672 ± 544

---

53.4

100

300

---

93.3

---

177.5

rat

1.22 ± 0.260

17.3 ± 0.981

---

4267 ± 1037

21,933 ± 4460

112,567 ± 35,842

0.250 ± 0.0

0.333 ± 0.144

0.417 ± 0.144

1.67 ± 0.577

0.5

---

5520 ± 999

57.1 ± 10.3

68.3 ± 10.1

126 ± 21.4

33,051 ± 4875

243,656 ± 41,452

200

minipig

16,100 ± 4990

91,300 ± 12,200

---

21,700

63,900

143,000

376,000

212,000

2.0

---

619,000

125

300

2.0

---

1,750,000

4,290,000

2.0

---

[¹⁴

C]-HST5040A was administered. Toxicokinetic data from GLP dose-range ﬁnding study in minipigs; values are the average of n = 12

minipigs; statistical analysis was not performed; values are day 21 values. Units are expressed as (ng eq/g); --- = not determined; AUC = area

‑inf

under the drug plasma concentration−time curve from time 0 extrapolated to inﬁnity; CL = clearance; C = maximum plasma concentration; F =

max

bioavailability; IV = intravenous; PO = oral; t = terminal elimination half-life; T = time of maximum plasma concentration; V = volume of

max

distribution.

042

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

kg). The [ C]-HST5040A-related material was quickly and

widely distributed throughout the body, with most tissues

showing the highest concentrations at 1 h post dose. The t₁_/₂’s

in plasma and blood were 2.4 and 2.5 h, respectively. The total

minipig, and human hepatocytes. [ C]-HST5040 had no

signiﬁcant impact on the hepatocyte viability in any species.

[ C]-HST5040 at 50 μM was relatively stable in rat, minipig,

and human hepatocytes, for which 86, 86, and 88% of total

radioactivity remained as unchanged HST5040 after the 4 h

incubation, respectively. An acyl glucuronide conjugate

(HST5040-glucuronide) was observed as the minor metabolite

in all species, accounting for <2% of total radioactivity in each

species after the 4 h incubation. Importantly, no unique human

metabolite was observed.

plasma and blood radioactivity decreased steadily after T_m_a_x

with both below the limit of quantitation (BQL) by 24 h post

dose. Overall, [ C]-HST5040A-derived radioactivity was

cleared rapidly from tissues in the rats, since at 24 h post

dose, most tissues had concentrations below the lower limit of

quantitation (LLOQ) and all tissues were BQL by 168 h after

dosing. Most tissues had tissue/plasma ratios lower than 1.00,

which indicates an overall low exposure in the tissues of male

rats relative to plasma. However, the liver/plasma ratio of

HST5040 was 4.6:1 at t = 1 h post dose, the highest ratio of

any tissue examined except the bile duct (14.5:1). This is

important because the liver is one of the key tissues for

mediating the desired pharmacology of HST5040. Plasma had

higher concentrations than blood at all time points post dose

Additionally, in vivo metabolite proﬁling experiments were

conducted in rats and minipigs. After a single dose of 10 mg/

kg (100 μCi/kg) [ C]-HST5040A, in addition to unchanged

HST5040, three metabolites were identiﬁed by LC/MS in

selected plasma, urine, and fecal samples of both species: the

HST5040-glucuronide that was observed in the in vitro

metabolism studies and two additional phase II metabolites,

identiﬁed as HST5040-carnitine and HST5040-glycine con-

jugates. The structures of all three metabolites were conﬁrmed

by comparison of the LC retention time, high-resolution full

scan MS, and MS/MS data with those of synthetically prepared

authentic standards. No phase I metabolites were observed in

either the in vitro or in vivo studies, supporting our original

hypothesis that HST5040 does not undergo β-oxidation and is

generally refractory to oxidative metabolism.

The potential for HST5040 to participate in PK-based

drug−drug interactions (DDIs) also was evaluated. DDI

studies assessed both the inhibitory and induction potential

of HST5040 against a broad panel of CYP450 isozymes and

human eﬄux and uptake transporters. HST5040 was tested

against human CYP1A2, CYP2B6, CYP2C8, CYP2C9,

CYP2C19, CYP2D6, and CYP3A4/5 using pooled, mixed

gender human liver microsomes. Both direct and time-

dependent inhibitions (TDIs) were evaluated. HST5040A at

concentrations of up to 10,000 μM showed no/limited direct

and TDI eﬀect (IC s > 10,000 μM) on all CYPs tested (Table

for all groups, which indicates that [ C]-HST5040 does not

have a high aﬃnity for the red blood cell (RBC) layer. The

highest concentrations of [ C]-HST5040A-derived radio-

activity were measured in the metabolic/excretory system, as

well as the gastrointestinal tract. The distribution to the brain

and spinal cord was measurable at 1 and 2 h, although lower

when compared to most other tissues, and was BQL by 8 and 4

h post dose, respectively. The main route of excretion for

[

(

C]-HST5040A-derived radioactivity was through urine

72−81%).

Similarly, the pharmacokinetics, absorption, distribution, and

excretion proﬁles of [ C]-HST5040A-derived radioactivity

were also evaluated in eight female Gottingen minipigs after a

single dose of 10 mg/kg (100 μCi/kg) [ C]-HST5040A (PK

data in Table 5). Following oral administration of [ C]-

HST5040A, radioactivity was quantiﬁable through at least 24 h

post dose in blood and plasma. Levels of radioactivity in blood

and plasma were BLQ in all animals by 96 and 72 h post dose,

respectively. Maximum concentrations of radioactivity oc-

curred at 1−2 h post dose. Overall, [ C]-HST5040A-derived

radioactivity was well absorbed. Urinary excretion was the

predominant route of elimination, accounting for a mean of

S6 ). In addition, HST5040A at concentrations of up to 4000

μM showed no induction potential for the mRNA expression

levels of CYP1A2, CYP2B6, and CYP3A4 in hepatocytes from

three diﬀerent donors (Table S7 ).

In addition, HST5040 was evaluated as an inhibitor of the

human ATP-binding cassette (ABC) transporters multidrug

resistance 1 (MDR1), breast cancer resistance protein

7.4% of the administered radioactive dose. Fecal excretion

was a minor elimination pathway, accounting for a mean of

.26% of the dose. Maximum concentrations of radioactivity

occurred at 2 h post dose for all tissues, and elimination of

radioactivity over time was evident in all tissues. Tissues with

the highest concentrations of radioactivity were the kidney,

bladder, liver, stomach, and ovaries. Tissue/plasma ratios

increased over time and were >1 for most tissues with

quantiﬁable radioactivity levels at 24 and 48 h post dose,

suggesting high partitioning of radioactivity into tissues and/or

slow elimination of radioactivity out of tissues. Exposure of

HST5040 was higher in the liver than any other tissue, with

liver/plasma ratios of 1.35:1, 20.0:1, 24.4:1, and 16.6: at t = 2,

(

BCRP), and bile salt export pump (BSEP) and the human

solute carrier (SLC) transporters MATE1, MATE2-K, OAT1,

OAT3, OATP1B1, OATP1B3, OCT1, and OCT2. The data

from these experiments are summarized in Table 6. HST5040

did not inhibit the BCRP-mediated E3S transport or MDR1-

mediated NMQ transport (up to 24,000 μM) and did not

inhibit BSEP-mediated taurocholate transport (up to 345 μM).

HST5040 also did not inhibit the MATE1- or MATE2-K-

mediated metformin transport (up to 14,900 μM) or OCT1 or

OCT2-mediated metformin transport (up to 14,900 μM).

HST5040 displayed very weak inhibition of OATP1B1-

mediated E217βG transport and OATP1B3-mediated CCK-8

transport (25.9 and 31.5%, respectively) at the maximum test

concentration of 14,900 μM and more potent inhibition of

OAT1-mediated tenofovir transport and OAT3-mediated E3S

transport (IC50 values of 3862 and 3511 μM, respectively).

The results of these DDI studies suggest that HST5040

displays minimal to no potential for signiﬁcant PK-based

drug−drug interaction in humans.

, 24, and 48 h post dose. [14C]-HST5040A-derived

radioactivity was present in the brain and spinal cord at 2

and 8 h post dose, with tissue/plasma ratios ranging between

.12 and 0.21, suggesting that some CNS penetration of

HST5040 occurs in minipigs. This may be important in the

potential of HST5040 to ameliorate the CNS-based toxicities

associated with PA and MMA.

Evaluation of the in vitro metabolic proﬁle of HST5040 and

identiﬁcation of the major HST5040 metabolites were carried

out by incubation of [ C]-HST5040A at 50 μM in rat,

043

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

Table 6. In Vitro Interaction Studies of HST5040 with

Human Transporter Proteins

While 2,2-dimethylbutanoic acid (HST5040) is a readily

available commercial compound, we could not identify any

vendors that currently manufacture this material to current

good manufacturing practice (cGMP) standards, so a simple

sodium salt formation of commercial-grade 2,2-dimethylbuta-

noic acid was not an appropriate option for manufacturing

HST5040A for clinical use. Therefore, we developed a cGMP

synthesis of HST5040A, the active pharmaceutical ingredient

HST5040 max conc.

maximum inhibition

(%)

(μM)

transporter

(μM)

MDR1

BCRP

BSEP

24,000

345

NIO

3862

3511

MATE1

14,900

(

API) for our human clinical trials. The synthetic approach we

MATE2-K

NIO

chose relies on the well-established dianion alkylation of

carboxylic acids. Indeed, the dianion of isobutyric acid has

OAT1

83.2 ± 3.5

90.9 ± 3.6

25.9

31.5

NIO

OAT3

been subjected to a number of electrophiles to provide α-

OATP1B1

OATP1B3

OCT1

29,30

quaternary carboxylic acid derivatives.

The synthesis is

described in Scheme 1. In brief, lithium diisopropylamide

OCT2

NIO

Scheme 1. Synthesis of HST5040A

Values are n = 1 except where indicated. Values are n = 2; NIO =

no interaction observed (<20%); NA = not applicable.

Since the data in Table 3 indicated that minipig pHeps were

the species most similar to human pHeps in their response to

HST5040, minipigs were selected as the nonrodent species for

toxicology studies. Because we anticipate clinical investigation

of HST5040 in a predominantly pediatric population, our

initial study was a 3 week good laboratory practice (GLP)-dose

range-ﬁnding study in neonatal minipigs. Groups of 12 Yucatan

minipigs, 2 day old at the ﬁrst dose, were given daily oral doses

of water or HST5040A at 20, 50, 125, or 300 mg/kg for 3

weeks at a dose volume of 5 mL/kg and then were euthanized

and necropsied.

Blood samples were taken from two minipigs/sex at intervals

after the ﬁrst, ﬁfth, tenth, and last doses and analyzed to

measure HST5040 concentration. Concentration data were

analyzed to derive toxicokinetic parameters; results after the

last dose are shown in Table 5. HST5040 was relatively rapidly

absorbed at all dose levels, with T_m_a_xvarying from 0.5 to 2 h

post dose. HST5040 was present in plasma for at least 24 h

after the ﬁrst and last doses. The relationship between the

HST5040A dose level and systemic exposure to HST5040 was

similar in both sexes. HST5040 exposure generally increased

slightly more than proportionally with the dose level. Exposure

increased between the ﬁrst and ﬁfth doses at all dose levels but

then did not change appreciably with continued dosing (data

not shown).

Reagents and conditions: (a) (1) LDA, THF/heptane/ethylbenzene,

°C to 40 °C then recooled to 5 °C; (2) bromoethane, 5 °C to rt;

and (3) H O and then HCl. (b) (1) MeONa/MeOH, MTBE and (2)

ACN, reﬂux. (c) ACN/H O 30:1, reﬂux, and then cooled to 0 °C.

(

LDA) is cooled to ca. 0 °C, and thencommercially available

isobutyric acid is added at a rate that maintains the

temperature below 5 °C. The resulting mixture is warmed to

ca. 40 °C to ensure complete anion formation and then

recooled to ca. 0 °C. Bromoethane is then added at a rate that

maintains the temperature below ca. 5 °C. The reaction is

warmed to rt and held until the reaction is complete.

Acidiﬁcation and workup provides 2,2-dimethylbutanoic acid

in good overall yield (75−95%) and >99% purity. Importantly,

this protocol allows for essentially complete conversion of the

starting material to the product, which minimizes the potential

need for further removal of residual isobutyric acid. The

sodium salt is then formed using sodium methoxide in

methanol in a methyl tert-butyl ether (MTBE)/ACN cosolvent

mixture. Subsequent recrystallization of crude HST5040A in

ACN/H O 30:1 aﬀords HST5040A as a white crystalline solid

in greater than 99.7% purity. This process has been utilized to

provide >35 kg of HST5040A. Detailed analytical character-

ization data for HST5040A can be found in Section S8.

Neonatal minipigs tolerated HST5040A at all dose levels in

the study. Only two ﬁndings were clearly related to HST5040

administration. One was minimally to moderately increased

vacuolation in renal tubules of all minipigs at ≥50 mg/kg/day,

with a dose-related increase in average grade. This ﬁnding was

characterized by an increase in both the number of cells with

cytoplasmic vacuoles and vacuole size in the epithelial cells of

tubules near the corticomedullary junction. Vacuolation was

not considered adverse because it was not accompanied by

evidence of renal tubule cell degeneration or necrosis or by

eﬀects on serum blood urea nitrogen, creatinine, or

phosphorus concentration. The second ﬁnding was minimally

increased mean serum albumin concentration and the resultant

increased albumin-to-globulin ratio at 300 mg/kg/day, which

was not considered adverse. Based on the absence of adverse

eﬀects, the 3 week no adverse event level (NOAEL) for

HST5040A in this study was 300 mg/kg/day, the highest dose

level tested, which is equivalent to an HST5040 dose level of

CONCLUSIONS

■

Herein, we have described the SAR of a series of low-

molecular-weight carboxylic acids with respect to their ability

to lower P-CoA levels in primary human hepatocytes derived

from patients with PA. We selected HST5040 (2,2-

dimethylbutanoic acid, compound 10) for further preclinical

evaluation based on its activity in this assay, its inability to

undergo canonical β-oxidation, and its early safety proﬁle in

human clinical studies for indications unrelated to our own.

We have also provided additional preclinical data in support

of HST5040 as potential therapeutic agents for treatment of

PA and MMA. HST5040 potently eﬀectively lowers the key

PA and MMA disease biomarkers P-CoA, M-CoA, MCA, and

C3/C2 ratio, with similar EC₅₀values for each biomarker.

Disease-related biomarkers in PA and MMA pHeps are

elevated ∼20-fold (P-CoA) to ∼40−80-fold (C3) compared

to normal pHeps. HST5040 reduces the levels of disease-

252 mg/kg/day.

related biomarkers to levels close to normal values, with P-CoA

044

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Article

within threefold and C3 within ninefold of normal hepatocyte

mm). The column conditions were 98% water with 0.05% TFA and

% MeOH to 100% MeOH over 5.5 min.

Materials. All carboxylic acids listed in Table 1 were purchased

levels. A putative working model of how HST5040 reduces

P-CoA and its derived metabolites involves the conversion of

HST5040 to HST5040-CoA which drives the redistribution of

free and conjugated CoA pools, resulting in the diﬀerential

reduction of the aberrantly high P-CoA and M-CoA. The

^f^o^r^m^a^tⁱ^oⁿ^o^f^H^S^T⁵⁰⁴⁰^-^C^o^A^o^c^c^u^r^s^a^t^sⁱ^mⁱ^l^a^r^E^C⁵0 to the

reduction in P-CoA. Analogues of HST5040 that did not form

a CoA metabolite were unable to reduce P-CoA, which further

supports the mechanistic link between HST5040-CoA and P-

CoA. The reduction of P-CoA and M-CoA, either by slowing

production [due to increased demands on the free CoA

from commercial sources and utilized without additional puriﬁcation.

All compounds were listed as ≥95% purity by the commercial

vendors.

Sodium 2,2-Dimethylbutanoic Acid (HST5040A). Step 1: A

solution of LDA (24% to 27%) in tetrahydrofuran (THF)/heptane/

ethylbenzene (approximately 2.5 equiv; 11.5 wt) is charged to a

reactor and cooled to 0−5 °C, and then, isobutyric acid (1.0 equiv;

1.0 wt) is added at a rate so as to maintain the reaction temperature

below 5 °C. The reaction mixture is warmed to between 35 and 45 °C

before recooling the solution to 0−5 °C. Bromoethane (approx-

imately 2.0 equiv; 2.5 wt) is charged to the reactor at a rate which

maintains the temperature below 5 °C, and then, the reaction is

warmed to between 15 and 25 °C and held until the reaction is

(

CoASH) pool] or enhancing clearance (to replenish the

CoASH pool), results in a net decrease in the CoA-derived

metabolites C3, MCA, and methylmalonic acid (MMA pHeps

complete. The reaction is cooled and H O (approximately 10

only).

volumes) is added at a rate so as to maintain the reaction temperature

<30 °C. The layers are separated, and the product-containing aqueous

phase is washed twice with MTBE. The pH of the product-containing

aqueous layer is adjusted to approximately pH = 1 by the addition of

HST5040A, the sodium salt of HST5040, displays excellent

solubility, passive permeability, and low plasma protein binding

across preclinical species. HST5040 displays minimal to no

potential for signiﬁcant drug−drug interaction risks in humans,

as judged by CYP inhibition and activation studies along with

drug transporter interaction studies. HST5040A also displays

good pharmacokinetic proﬁles in mice, rats, and minipigs,

showing rapid absorption, low to moderate clearance, and

moderate to high oral bioavailability with elimination kinetics

that appear suitable for QD or BID dosing. In QWBA studies,

HST5040A is broadly distributed into all tissues, including

some CNS penetration, and shows a steady decrease in tissue

levels over time with elimination primarily via renal excretion.

In vitro and in vivo metabolite proﬁling and identiﬁcation

studies demonstrate that HST5040 undergoes minimal

metabolism in all species, with the major metabolite being

the HST5040-glucuronide adduct along with minor metabo-

lites being HST5040-glycine and HST5040-carnitine adducts.

Importantly, no human-speciﬁc metabolites were identiﬁed in

these studies. Finally, HST5040A was well-tolerated in a 3

week dose-range-ﬁnding toxicity study in neonatal minipigs,

the species with the most similar on-target pharmacology to

humans. Based on the absence of adverse eﬀects, the NOAEL

for HST5040A was 300 mg/kg/day, the highest dose level

tested, which provided exposures well above those projected to

occur at the clinical eﬃcacious dose. Overall, these data are

similar to previously reported preclinical safety data with this

N aqueous hydrochloric acid (HCl) while maintaining the

temperature <30 °C. The product-containing acidic aqueous phase

is extracted three times with MTBE to provide a solution of HST5040

in the organic phase. The product-containing organic phase is washed

twice with H O, and then, a sample is removed for determination of

isobutyric acid content. If the solution contains more than 0.1%

isobutyric acid, additional H O washes are conducted. Once the level

of isobutyric acid is <0.1%, MTBE is added and the solvent is

evaporated to a minimum. An additional charge of MTBE is added,

and the additional solvent is evaporated to aﬀord a solution of

HST5040 in MTBE (about 50% wt/wt). The product is carried on to

step 2 (expected yield 75−95% theory) without further isolation.

Step 2: To a solution of HST5040 in MTBE (1 weight, about 50%

wt/wt in solution) is added approximately 2 volumes of MTBE. The

solution is cooled to 0−5 °C, and a solution of sodium methoxide in

methanol (25% wt/wt, 0.95 equiv, 1.77 weights) is added while

maintaining the temperature below 20 °C. Acetonitrile (ACN,

approximately 10 volumes) is added, and the contents of the reactor

are distilled to a minimum volume before an additional charge of

ACN (about 5 volumes) is made. The contents of the reactor are

again distilled to a minimum volume before ACN (about 6 volumes)

and water (about 0.45 volumes) are added. The suspension is heated

to reﬂux at which point the solids dissolve. The solution is slowly

cooled to approximately 0 °C during which time the product

crystallizes. The solids are ﬁltered and washed with ACN and then

dried in vacuo at <60 °C to aﬀord intermediate-grade HST5040A

(

expected yield 79−100% theory).

1,32

molecule.

Based on these data and the pharmacological

Step 3: A suspension of HST5040A (1 weight) in approximately 10

activity in disease models, HST5040 was selected as a clinical

development candidate for the treatment of PM and MMA.

volumes of ACN and about 0.3 volumes of water is warmed to reﬂux

and a solution is obtained. If necessary, additional aliquots of 0.05

volumes of water may be added portionwise until a solution is

obtained. The solution is slowly cooled to approximately 0 °C, and

the solids are ﬁltered and washed with ACN. The puriﬁed HST5040A

EXPERIMENTAL SECTION

■

General Procedures. All solvents and chemicals were used as

purchased without further puriﬁcation. Unless otherwise indicated, all

temperatures are expressed in °C (degrees Centigrade). H NMR

is dried in vacuo at <60 °C to a constant weight (expected yield 75−

95% theory). H NMR (500 MHz, DMSO-d ): δ 1.47 (q, J = 9.5,

2H), 0.94 (s, 6H), 0.77 (t, J = 9.5, 3H). C NMR (125 MHz, D O):

spectra were recorded on a Varian VXR-400 or a Varian Unity-400 at

δ 188.1, 43.8, 33.4, 25.1, 9.7. HRMS (ESI): m/z for C H O calcd

00 MHz ﬁeld strength. Chemical shifts are expressed in parts per

million (ppm, δ units). Coupling constants (J) are in units of hertz

Hz). Splitting patterns describe apparent multiplicities and are

designated as s (singlet), d (doublet), t (triplet), q (quartet), m

multiplet), quin (quintet), or br (broad). High-resolution mass

115.080; found, 115.076. Elemental analysis for C H NaO : Calcd

(%): C 52.17, H 8.03; Found 52.06, H 7.15.

Primary Hepatocyte Static Cell Culture Experiments.

6 11 2

(

Primary human hepatocytes were isolated from liver tissue from

(

PA, MMA, and normal donors as previously described. Primary

spectrometry (HRMS) was performed on a Thermo Fisher Orbitrap

Velos Pro high-resolution mass spectrometer. The sample was

analyzed by electrospray ionization in the negative ion mode at a

mass resolution setting of 15,000. The sample was delivered to the

instrument via direct infusion in a methanol/water solution. The LC/

MS was run using a C-18 reverse phase column (2.1 ID, 3.5 μm, 50

human hepatocytes from individual donors were plated in a collagen

gel sandwich conﬁguration on 48-well plates (Corning) and cultured

in customized hepatocyte maintenance medium purchased from

Corning using previously described protocols. Primary hepatocytes

were pretreated with HST5040 for 30 min, followed by a 1 h

incubation with 13C-isoleucine. Hepatocytes were harvested and cell

045

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Complete contact information is available at:

Article

lysates were enriched for the quantitation for acyl-CoAs by MS/MS as

previously described.

John E. Reardon − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Mouse, rat, dog, minipig, cynomolgus monkey, and human

hepatocytes were purchased as preseeded Ready-Plates from Sekisui

XenoTech (Cambridge, MA). Opti Culture hepatocyte media was

also purchased from XenoTech (K8400) for the feeding and

treatment of the cultured hepatocytes. Cells were treated with

HST5040 for 1 h and harvested, and cell lysates were enriched for the

https://pubs.acs.org/10.1021/acs.jmedchem.1c00124

Author Contributions

All authors have given approval to the ﬁnal version of the

manuscript.

quantitation for acyl-CoAs by MS/MS as previously described.

Animal Studies. Studies in rats and minipigs were conducted at

contract research organizations (CROs) that specialize in such

studies. All studies were conducted in compliance with each CRO’s

standard operating procedures (SOPs). With respect to the welfare of

animals used in these studies, the animals were maintained and used

as recommended in the Guide for the Care and Use of Laboratory

Animals (Institute of Laboratory Animal Resources, National

Academy Press, Washington, D.C., 20011), and each CRO is

regularly inspected by the USDA for compliance with animal welfare

regulations.

Funding

This research was supported in part by SBIR grant

R44TR001407 and HemoShear Therapeutics, Inc.

Notes

The authors declare the following competing ﬁnancial

interest(s): The following authors have units of ownership in

HemoShear Therapeutics, Inc.: Allison J. Armstrong, Brad R.

Henke, Maria Sol Collado, Matthew W. Olson, Stephen A.

Hoang, Justin M. Taylor, Brian A. Johns, Brian R. Wamhoﬀ,

John E. Reardon, and Robert A Figler.

ASSOCIATED CONTENT

sı Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.jmedchem.1c00124.

■

ACKNOWLEDGMENTS

■

The authors would like to express thanks to Andrew Pryor,

Nathan Day, Sarah Mayer, Christin Hamilton, and Crystal

Passmore for technical assistance. The authors thank Wuxi

Apptec (Cranberry, NJ) and Sinclair Research Center

(Auxvasse, MO) for their technical assistance. The authors

thank Chuck Poole for guidance on the studies and Mark

Hayes for critical review of the manuscript.

MFS spreadsheet (CSV)

Concentration−response curves for active compounds in

Table 1, details on calculation of adjusted AUDRC,

summary plot of normalized adjusted AUDRC for

compounds in Table 1, concentration−response curves

for CoA formation of selected compounds, example

concentration−response curves for HST5040 in the

HemoShear hepatocyte system, CYP activity data for

HST5040A, and full analytical characterization data for

HST5040A (PDF)

ABBREVIATIONS

■

ABC, ATP-binding cassette; ACN, acetonitrile; ADME,

absorption, distribution, metabolism, and excretion; API,

active pharmaceutical ingredient; AUDRC, area under the

dose−response curve; BCRP, breast cancer resistance protein;

BID, twice daily; BQL, below the limit of quantitation; BSEP,

bile salt export pump; C2, acetyl carnitine; C3, propionyl

carnitine; cGMP, current good manufacturing practice; cLogP,

calculated LogP; CNHS, College of Nursing and Health

Sciences; CNS, central nervous system; CoASH, coenzyme A;

CYP, cytochrome P450; DDI, drug−drug interaction; EC,

eﬀective concentration; ER, extraction ratio; GI, gastro-

intestinal; GLP, good laboratory practice; GMP, good

manufacturing practice; HCl, hydrochloric acid; HPLC, high-

performance liquid chromatography; HR/MS, high-resolution

mass spectrometry; IC, inhibitory concentration; IND,

investigational new drug; IRB, institutional review board; IV,

intravenous; kg, kilogram; LC/MS, liquid chromatography/

mass spectrometry; LDA, lithium diisopropylamide; LLOQ,

lower limit of quantitation; LSC, liquid scintillation counter;

MATE, multi-antimicrobial extrusion protein; MCA, 2-

methylcitric acid; M-CoA, methylmalonyl coenzyme A ester;

MDCKII, Madin−Darby canine kidney II cell line; MDR,

multidrug resistance; MMA, methylmalonic acidemia; MTBE,

methyl tert-butyl ether; MUT, methylmalonyl-CoA mutase;

MW, molecular weight; NOAEL, no adverse event level; OAT,

organic anion transporter; OCT, organic cation transporter;

PA, propionic acidemia; PCC, propionyl-CoA carboxylase; P-

CoA, propionyl coenzyme A ester; pHeps, primary hepato-

cytes; PK, pharmacokinetics; PO, oral; PSA, polar surface area;

QD, once daily; QWBA, quantitative whole-body auto-

radiography; SAR, structure−activity relationship; SLC, solute

carrier; RBC, red blood cell; TCA, tricarboxylic acid cycle;

TDI, time-dependent inhibition

AUTHOR INFORMATION

Corresponding Author

Brian A. Johns − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States; orcid.org/

■

000-0001-5526-3682 ; Email: johns@hemoshear.com

Authors

Allison J. Armstrong − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Brad R. Henke − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Maria Sol Collado − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Justin M. Taylor − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Taylor D. Pourtaheri − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

John E. Dillberger − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Thomas D. Roper − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Brian R. Wamhoﬀ − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Matthew W. Olson − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Robert A. Figler − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

Stephen A. Hoang − HemoShear Therapeutics Inc.,

Charlottesville, Virginia 22902, United States

046

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

Case, S.; Keefer, J. R.; Faller, D. V.; Welch, W. C.; Berenson, R. J.

Article

(

15) Perrine, S. P.; Wargin, W. A.; Boosalis, M. S.; Wallis, W. J.;

REFERENCES

■

Evaluation of Safety and Pharmacokinetics of Sodium 2,2

Dimethylbutyrate, a Novel Short Chain Fatty Acid Derivative, in a

Phase 1, Double-Blind, Placebo-Controlled, Single- and Repeat-dose

Studies in Healthy Volunteers. J. Clin. Pharmacol. 2011, 51, 1186−

1) Fraser, J. L.; Venditti, C. P. Methylmalonic and propionic

acidemias: clinical management update. Curr. Opin. Pediatr. 2016, 28,

(

82−693.

2) Chapman, K. A.; Gropman, A.; MacLeod, E.; Stagni, K.;

Summar, M. L.; Ueda, K.; Mew, N. A.; Franks, J.; Island, E.; Matern,

D.; Pena, L.; Smith, B.; Sutton, V. R.; Urv, T.; Venditti, C.;

Chakrapani, A. Acute management of propionic academia. Mol. Genet.

Metab. 2012, 105, 16−25.

3) Baumgartner, M. R.; Hörster, F.; Dionisi-Vici, C.; Haliloglu, G.;

Karall, D.; Chapman, K. A.; Huemer, M.; Hochuli, M.; Assoun, M.;

Ballhausen, D.; Burlina, A.; Fowler, B.; Grunert, S. C.; Grunewald, S.;

Honzik, T.; Merinero, B.; Pérez-Cerdá, C.; Scholl-Burgi, S.; Skovby,

F.; Wijburg, F.; MacDonald, A.; Martinelli, D.; Sass, J. O.;

Valayannopoulos, V.; Chakrapani, A. Proposed guidelines for the

diagnosis and management of methylmalonic and propionic academia.

Orphanet J. Rare Dis. 2014, 9, 130.

1194.

(

16) Kutlar, A.; Ataga, K.; Reid, M.; Vichinsky, E. P.; Neumayr, L.;

Blair-Britt, L.; Labotka, R.; Glass, J.; Keefer, J. R.; Wargin, W. A.;

Berenson, R.; Perrine, S. P. A Phase 1/2 Trial of HQK-1001, an Oral

Fetal Globin Inducer, in Sickle Cell Disease. Am. J. Hematol. 2012, 87,

(

017−1021.

17) Fucharoen, S.; Inati, A.; Siritanaratku, N.; Thein, S. L.; Wargin,

W. C.; Koussa, S.; Taher, A.; Chaneim, N.; Boosalis, M.; Berenson, R.;

Perrine, S. P. A randomized phase I/II trial of HQK-1001, an oral

fetal globin gene inducer, in β -thalassaemia intermedia and HbE/β -

thalassaemia. Br. J. Haematol. 2013, 161, 587−593.

(

18) Kutlar, A.; Reid, M. E.; Inati, A.; Taher, A. T.; Abboud, M. R.;

(

4) Grunert, S. C.; Muellerleile, S.; Silva, L.; Barth, M.; Walter, M.;

El-Beshlawy, A.; Buchanan, G. R.; Smith, H.; Ataga, K. I.; Perrine, S.

P.; Ghalie, R. G. A dose-escalation phase IIa study of 2,2-

dimethylbutyrate (HQK-1001), an oral fetal globin inducer, in sickle

cell disease. Am. J. Hematol. 2013, 88, E255−E260.

Walter, K.; Meissner, T.; Lindner, M.; Ensenauer, R.; Santer, R.;

Bodamer, O. A.; Baumgartner, M. R.; Brunner-Krainz, M.; Karall, D.;

Haase, C.; Knerr, I.; Marquardt, T.; Hennermann, J. B.; Steinfeld, R.;

Beblo, S.; Koch, H.-G.; Konstantopoulou, V.; Scholl-Burgi, S.; van

(19) Reid, M. E.; El Beshlawy, A.; Inati, A.; Kutlar, A.; Abboud, M.

Teeffelen-Heithoff, A.; Suormala, T.; Sperl, W.; Kraus, J. P.; Superti-

Furga, A.; Schwab, K. O.; Sass, J. O. Propionic acidemia: neonatal

versus selective metabolic screening. J. Inherited Metab. Dis. 2012, 35,

R.; Haynes, J.; Ward, R.; Sharon, B.; Taher, A. T.; Smith, W.;

Manwani, D.; Ghalie, R. G. A double-blind, placebo-controlled phase

II study of the efficacy and safety of 2,2-dimethylbutyrate (HQK-

(

1−49.

001), an oral fetal globin inducer, in sickle cell disease. Am. J.

Hematol. 2014, 89, 709−713.

20) Dash, A.; Simmers, M. B.; Deering, T. G.; Berry, D. J.; Feaver,

5) Grunert, S. C.; Muellerleile, S.; Silva, L.; Barth, M.; Walter, M.;

Walter, K.; Meissner, T.; Lindner, M.; Ensenauer, R.; Santer, R.;

Bodamer, O. A.; Baumgartner, M. R.; Brunner-Krainz, M.; Karall, D.;

Haase, C.; Knerr, I.; Marquardt, T.; Hennermann, J. B.; Steinfeld, R.;

(

R. E.; Hastings, N. E.; Pruett, T. L.; LeCluyse, E. L.; Blackman, B. R.;

Wamhoff, B. R. Hemodynamic flow improves rat hepatocyte

morphology, function, and metabolic activity in vitro. Am. J. Physiol.:

Cell Physiol. 2013, 304, C1053−C1063.

Beblo, S.; Koch, H.-G.; Konstantopoulou, V.; Scholl-Burgi, S.; van

Teeffelen-Heithoff, A.; Suormala, T.; Sperl, W.; Kraus, J. P.; Superti-

Furga, A.; Schwab, K. O.; Sass, J. O. Propionic acidemia: clinical

course and outcome in 55 pediatric and adolescent patients. Orphanet

J. Rare Dis. 2013, 8, 6.

(21) Feaver, R. E.; Cole, B. K.; Lawson, M. J.; Hoang, S. A.;

Marukian, S.; Blackman, B. R.; Figler, R. A.; Sanyal, A. J.; Wamhoff, B.

R.; Dash, A. Development of an in vitro human liver system for

interrogating nonalcoholic steatohepatitis. JCI Insight 2016, 1,

No. e90954.

Kong, L.-j.; Mackman, N.; Blackman, B. R.; Wamhoff, B. R.

6) Skonberg, C.; Olsen, J.; Madsen, K. G.; Hansen, S. H.; Grillo, M.

P. Metabolic activation of carboxylic acids. Expert Opin. Drug Metab.

Toxicol. 2008, 4, 425−438.

7) Collado, M. S.; Armstrong, A. J.; Olson, M.; Hoang, S. A.; Day,

(22) Simmers, M. B.; Cole, B. K.; Ogletree, M. L.; Chen, Z.; Xu, Y.;

(

Hemodynamics associated with atrial fibrillation directly alters

thrombotic potential of endothelial cells. Thromb. Res. 2016, 143,

N.; Summar, M.; Chapman, K. A.; Reardon, J.; Figler, R. A.; Wamhoff,

B. R. Biochemical and anaplerotic applications of in vitro models of

propionic acidemia and methylmalonic acidemia using patient-derived

primary hepatocytes. Mol. Genet. Metab. 2020, 130, 183−196.

(

4−39.

23) Cole, B. K.; Simmers, M. B.; Feaver, R.; Qualls, C. W., Jr.;

(

8) Schomburg, I.; Chang, A.; Ebeling, C.; Gremse, M.; Heldt, C.;

Collado, M. S.; Berzin, E.; Figler, R. A.; Pryor, A. W., Jr.; Lawson, M.;

Mackey, A.; Manka, D.; Wamhoff, B. R.; Turk, J. R.; Blackman, B. R.

An in vitro cynomolgus vascular surrogate system for preclinical drug

assessment and human translation. Arterioscler., Thromb., Vasc. Biol.

Huhn, G.; Schomburg, D. BRENDA, the enzyme database: updates

and major new developments. Nucleic Acids Res. 2004, 32, D431−

D433.

9) Beam, A.; Motsinger-Reif, A. Beyond IC s: Towards Robust

Statistical Methods for in vitro Association Studies. J. Pharmacoge-

nomics Pharmacoproteomics 2014, 5, 1000121.

Summar, M. L.; Wamhoff, B. R.; Dash, A. Recapitulation of metabolic

015, 35, 2185−2195.

24) Chapman, K. A.; Collado, M. S.; Figler, R. A.; Hoang, S. A.;

Armstrong, A. J.; Cui, W.; Purdy, M.; Simmers, M. B.; Yazigi, N. A.;

defects in a model of propionic acidemia using patient-derived

primary hepatocytes. Mol. Genet. Metab. 2016, 117, 355−362.

(25) Armstrong, A. J.; Collado, M. S.; Henke, B. R.; Olson, M. W.;

Hoang, S. A.; Hamilton, C. A.; Pourtaheri, T. D.; Chapman, K. A.;

Summar, M. M.; Johns, B. A.; Wamhoff, B. A.; Reardon, J. A.; Figler,

R. A. A Novel Small Molecule Approach for the Treatment of

Propionic and Methylmalonic Acidemias. Mol. Genet. Metab. 2021,

DOI: 10.1016/j.ymgme.2021.03.001. , in press

10) Hsieh, J.-H.; Sedykh, A.; Huang, R.; Xia, M.; Tice, R. R. A data

analysis pipeline accounting for artifacts in Tox21 quantitative high-

throughput screening assays. J. Biomol. Screen. 2015, 20, 887−897.

(

Reif, A. A. A comparison of association methods for cytotoxicity

mapping in pharmacogenomics. Front. Genet. 2011, 2, 86.

Hyperammonaemia in classic organic acidaemias: a review of the

literature and two case histories. Orphanet J. Rare Dis. 2018, 13, 219.

(

11) Brown, C.; Havener, T. M.; Everitt, L.; McLeod, H.; Motsinger-

12) Häberle, J.; Chakrapani, A.; Ah Mew, N.; Longo, N.

13) de las Heras, J.; Aldámiz-Echevarría, L.; Martínez-Chantar, M.-

L.; Delgado, T. C. An update on the use of benzoate, phenylacetate

and phenylbutyrate ammonia scavengers for interrogating and

modifying liver nitrogen metabolism and its implications in urea

cycle disorders and liver disease. Expert Opin. Drug Metab. Toxicol.

(26) Guenzel, A. J.; Hofherr, S. E.; Hillestad, M.; Barry, M.; Weaver,

E.; Venezia, S.; Kraus, J. P.; Matern, D.; Barry, M. A. Generation of a

Hypomorphic Model of Propionic Acidemia Amenable to Gene

Therapy Testing. Mol. Ther. 2013, 21, 1316−1323.

(27) Forny, P.; Schumann, A.; Mustedanagic, M.; Mathis, D.; Wulf,

M.-A.; Nägele, N.; Langhans, C.-D.; Zhakupova, A.; Heeren, J.;

Scheja, L.; Fingerhut, R.; Peters, H. L.; Hornemann, T.; Thony, B.;

Kölker, S.; Burda, P.; Froese, D. S.; Devuyst, O.; Baumgartner, M. R.

Novel mouse models of methylmalonic aciduria recapitulate

14) Surendradoss, J.; Chang, T. K. H.; Abbott, F. S. Assessment of

017, 13, 439−448.

the role of in situ generated (E)-2,4-diene-valproic acid in the toxicity

of valproic acid and (E)-2-ene-valproic acid in sandwich-cultured rat

hepatocytes. Toxicol. Appl. Pharmacol. 2012, 264, 413−422.

047

J. Med. Chem. 2021, 64, 5037−5048

Journal of Medicinal Chemistry

28) For a review, see: Petragnani, N.; Yonashiro, M. The reations of

Article

phenotypic traits with a genetic dosage effect. J. Biol. Chem. 2016, 291,

0563−20573.

dianions of carboxylic acids and ester enolates. Synthesis 1982, 1982,

21−578.

29) Creger, P. L. Metalated carboxylic acids. I. Alkylation. J. Am.

Chem. Soc. 1967, 89, 2500−2501.

30) Crimmins, M. T.; Mascarella, S. W.; DeLoach, J. A. Conjugate

Addition-Cycloacylation of Acetylenic Diesters: Synthesis of Highly

Functionalized Cyclopentenones. J. Org. Chem. 1984, 49, 3033−3035.

(31) Terse, P. S.; Johnson, J. D.; Hawk, M. A.; Ritchie, G. D.; Ryan,

M. J.; Vasconcelos, D. Y.; Contos, D. A.; Perrine, S. P.; Peggins, J. O.;

Tomaszewski, J. E. Short-term Toxicity Study of ST-20 (NSC-

(

41804) by Oral Gavage in Sprague-Dawley Rats. Toxicol. Pathol.

011, 39, 614−622.

32) Perrine, S. P.; Faller, D. V.; Berenson, R. J. Methods and low

dose regimens for treating red blood cell disorders. U.S. Patent

20,110,251,149 A1, October 13, 2011.

048