The Antileishmanial Activity of Isoforms 6- and 8-Selective Histone Deacetylase Inhibitors

Abstract

Histone deacetylase inhibitors (HDACi) pleiotropy is largely due to their non-selective inhibition of various cellular HDAC isoforms. Connecting inhibition of a specific isoform to biological responses and/or phenotypes is essential toward deconvoluting HDACi pleiotropy. The contribution of classes I and II HDACs to the antileishmanial activity of HDACi was investigated using the amastigote and promastigote forms of L. donovani. We observed that the antileishmanial activities of HDACi are largely due to the inhibition of HDAC6-like activity. This observation could facilitate the development of HDACi as antileishmanial agents.

Leishmaniasis is a parasitic disease caused by protozoan parasites of the genus Leishmania. Human leishmaniases are present on all continents, but in Antarctica and Australia.¹ Although the majority of leishmanial infections occur in the developing world, leishmaniasis is increasingly common in developed nations as well.^2,3 Globally, leishmaniasis is endemic in more than eighty countries and affects about 2 million people per year, with approximately 350 million individuals at risk of infection.⁴

The limited chemotherapeutic options for the treatment of leishmanial infections are based on amphotericin B, miltefosine, paromomycin, and antimony agents such as sodium stibogluconate and meglumine antimoniate. These are unsatisfactory agents because they require a long course of parenteral injection and many suffer from unacceptable toxicity and poor efficacy.^5–8 Therefore, there is an urgent need for affordable and safe alternative agents to curtail leishmaniasis.

Leishmania is a member of trypanosomatid protozoa, which are rapidly growing eukaryotes. The genomes of Leishmania, trypanosomatids and many human pathogenic protozoans contain several genes encoding for different histone deacetylase (HDAC) isozymes.^9–16 In the Leshmania genome, three genes putatively encoding for NAD⁺-dependent HDACs and four genes encoding for zinc-dependent HDACs have been identified. Further analysis of the genes encoding for zinc-dependent HDACs confirmed the presence of major conserved domains and identified these as HDAC1, RpdA/Rpd3 (both Class I), HDAC5 and HDAC4/6 (both class II) homologs (Unpublished results).

HDACs are the primary cellular lysine deacetylases,¹⁷ and their inhibition elicits an array of phenotypes including anti-proliferative activity against nearly all types of transformed cells. Two HDAC inhibitors (HDACi), suberoylanilide hydroxamic acid (SAHA) and FK-228 (romidepsin) (Figure 1), are currently approved by the FDA for the treatment of cutaneous T-Cell lymphoma (CTCL).¹⁷ Interestingly, Leishmania are also susceptible to HDACi,^18–20 suggesting that HDAC isozymes are essential for their viability and proliferation. However, the identity of the specific Leishmania HDAC isozymes, which confer antileishmanial activity to HDACi is currently unknown. Moreover, the structure activity relationship (SAR) of the antileishmanial activity of HDACi has not been completely worked out.

FIGURE 1.

Representative small molecules HDAC inhibitors.

The antiprotozoal activity of HDACi has been suggested to result from the inhibition of protozoan’s proteins similar to the class I family HDACs. This suggestion has been supported in the case of Plasmodium falciparum, the causative organism of the most prevalent form of human malaria, with the identification and characterization of PfHDAC-1 which has homology with the class I family of HDACs from human, chicken, frog and Saccharomyces cerevisiae.^13,20,21 Additionally, the antimalarial activities of most HDACi tracked very well with their anti- PfHDAC-1 activity.^18,19,22,23 However, the correlation between class I family HDAC inhibition and antileishmanial activity is a bit tenuous. The pan-HDAC inhibitor, SAHA, is potently active against P. falciparum but more than 70-fold less active against L. donovani, the causative organism of visceral leishmaniasis.^18–20 Moreover, several HDACi have inverse relation between HDAC1 inhibition and antileishmanial activity.²⁰ The disparities in the sensitivity of protozoans such as P. falciparum and L. donovani to HDACi may be due to differences in the key amino acid residues within the active sites of their class I HDAC isozymes. Alternatively, HDAC sub-classes other than class I may be essential for the survival and proliferation, particularly for Leishmania parasites and may be the potential target for action of antileishmanial HDACi.

We have previously identified a new class of non-hydroxamate HDACi derived from 3-hydroxypyridine-2-thione (3HPT).²⁴ We observed that aryl- and diaryl-analogs of these 3HPT-derived HDACi have selective inhibitory activity against HDAC6 or HDAC8 but are otherwise inactive against HDAC1. We envisioned that these 3HPT-derived compounds could constitute useful molecular probes for parsing out the contribution of inhibition of classes I and II HDACs to the antileishmanial activity of HDACi. Herein, we showed that despite their inactivity against HDAC1, these 3HPT-HDACi potently inhibit the viability of the amastigote and promastigote forms of L. donovani. Additionally, we provided evidence which suggested that the antileishmanial activity of these and other HDACi could be largely due to inhibition of HDAC6-like activity.

We initially evaluated the in vitro antileishmanial activities of the 3HPT-derived HDACi and their corresponding 3-hydroxypyridin-2-one (3HP) analogs against the amastigote and promastigote stages of L. donovani, the two different stages of Leishmania species. The amastigote is found thriving in the macrophages of the mammalian host while the promastigote starts as the non-infective stage, which replicates within the insect vector and subsequently transforms into infective metacyclic promastigote.¹¹ Viability of the two stages of L. donovani was determined using the standard Alamar blue assay, modified to a fluorometric assay.²⁵ We used suberoylanilide hydroxamic acid (SAHA), a standard HDACi as well as Amphotericin B and pentamidine, standard antileishmanial agents, as positive controls. We observed that for each matched pair, the 3HP compounds are relatively weakly cytotoxic to the promastigote form while the 3HPT-HDACi compounds are potently cytotoxic (Table 1). The lead 3HP compound 1a is inactive at the maximum concentration tested (40 µg/mL) while its 3HPT analog, 1b, is weakly cytotoxic to the promastigote stage of L. donovani. Subsequent modification to 1a and 1b, as described in our previous publication,²⁴ afforded matched pairs 2a/2b through 13a/13b. Without exception, the 3HPT analogs of each pair (analog “b”) are much more cytotoxic than their 3HP congeners with IC₅₀ ranging from 100 ng/mL to single digit µg/mL. Overall, the potency enhancement of the 3HPT compounds relative to their 3HP analogs range from 11- to >90-folds. The most potent analogs are compounds 2b, 3b, 5b and 7b–10b (IC₅₀ 100–400ng/mL). The relative inactivity of the 3HP compounds (analog “a”) may be due to their lack of HDAC inhibition activity. Conversely, all the 3HPT compounds, except 5b and 10b–12b, inhibit both HDAC6 and HDAC8. Compounds 5b, 10b and 12b inhibit HDCA8 while 11b inhibits HDAC6. For this set of compounds, we did not see any clear evidence suggesting that the preferential inhibition of HDAC6 or HDAC8 is more deleterious to the viability of the promastigote stage of L. donovani. Rather, our result seems to suggest that HDAC1 inhibition may not be necessary for the cytotoxicity of HDACi against the L. donovani promastigote stage.

Table 1.

In Vitro HDAC inhibition (nM) and Antileishmanial Activities (µg/mL)

Compound		Promastigote IC50 (µg/mL)*	Amastigote IC50 (µg/mL)*	HDAC6 IC50 (nM)a	HDAC8 IC50 (nM)a
	X
	1a	>40	>40	NI	NI
	1b	12.7	30.7	681 ± 110	3675 ± 1201
	2a	9.3	>40	NI	NI
	2b	0.1	6.5	457 ± 27	1272 ± 200
	3a	6.8	30.9	NI	NI
	3b	0.1	6.3	847 ± 188	4283 ± 1548
	4a	4.6	5	NI	NI
	4b	0.7	6.1	957 ± 159	2075 ± 459
	5a	9	>40	NI	NI
	5b	0.4	11.2	44%†	1701 ± 717
	6a	26.7	>40	NI	NI
	6b	0.9	31.7	372 ± 35	1907 ± 771
	7a	6	29	NI	NI
	7b	0.3	31.3	454 ± 42	800 ± 304
	8a	12.6	32.5	NI	NI
	8b	0.3	32.6	812 ± 286	2496 ± 1180
	9a	27.6	31.6	NI	NI
	9b	0.3	34.4	306 ± 69	3105 ± 1649
	10a	7.5	4	NI	NI
	10b	0.2	7.4	NI	2858 ± 944
	11a	38.9	>40	NI	NI
	11b	2.3	6.1	2390 ± 458	NI
	12a	>40	18	NI	NI
	12b	3.2	6.2	41%†	1570 ± 1067
	13a	14.5	>40	NI	NI
	13b	1.3	5.3	1023 ± 99	1868 ± 723
SAHA	-	25.7	>40	144 ± 23	232 ± 19
Tubastatin A	-	4.4	8.8	15 ± 1b	854 ± 40b
PCI-34051	-	30.5	>40	2900c	10c
Pentamidine	-	0.6	8.9	NT	NT
Amphotericin B	-	0.4	1.6	NT	NT

NT = Not tested; NI = No Inhibition at maximum tested concentration of at 10 µM (percentage inhibition < 20%);

^*Average ± SD of triplicate observations.

^†% inhibition of the compounds at 10 µM are given if the IC₅₀ was above 10 µM;

^aCited from ref. 24, except for the control Tubastatin A and PCI-34051.

^bCited from ref. 26.

^cCited from ref. 27.

To further delineate the contribution of HDAC6- and HDAC8-like activities to the viability of the promastigote stage, we investigated the effects of exposure to Tubastatin A²⁶ and PCI-34051,²⁷ prototypical HDAC6 and HDAC8 selective HDACi respectively, on cell viability. Similar to the pan HDACi SAHA, PCI-34051 has only but minimal effect on cell viability. Interestingly, Tubastatin A is much more cytotoxic to the promastigote stage of L. donovani than PCI-34051 with IC₅₀ of 4.4µg/mL. This data suggests that the inhibition of the HDAC6-like activity is more deleterious to the viability L. donovani promastigote stage. The fact that the apparently HDAC8- selective compounds 5b and 10b maintained potent antileishmanial activity suggests that their cytotoxicity may be due to perturbation of other as yet to be identified intracellular targets.

The axenic amastigote form is generally less responsive to drug treatment including the standard antileishmanial agents, Amphotericin B and pentamidine, and all HDACi investigated. All 3HP compounds are virtually nontoxic to the axenic amastigote except 4a and 10a, which are about equipotent to both stages of L. donovani. Relative to their effects on the promastigote stage, the 3HPT compounds are between 2- to 115-fold less active against the promasitgote stage. Nevertheless, compounds 2b–4b and 10b–13b have single digit µg/mL cytotoxic activity. The pan HDACi SAHA and HDAC8-selective HDACi are non-toxic to the amastigote stage at maximum tested concentration of 40µg/mL. Interestingly, the cytotoxic activity of the HDAC6-selective inhibitor Tubastatin A is only 2-fold lower than that of its effect on the promastigote stage (IC₅₀ of 8.8µg/mL). These data suggest that HDAC6-like activity is essential for the viability of the amastigote stage of L. donovani as well. It is worth noting here that a stage-specific response of Leishmania spp to HDACi has been previously observed and attributed to overexpression of SIR2, a cytoplasmic NAD⁺-dependent HDAC.²⁸ The weaker response of the amastigote stage to the active HDACi described herein could also be due to compensation from the upregulated SIR2 activity.

To investigate the activity or lack thereof of these 3HPT HDACi against the therapeutically relevant mammalian host stage of L. donovani, we tested all compounds in vitro in amastigote-macrophage assay. We used a human THP1 macrophage cell line both as the amastigote host cell and as a control for the determination of drug selective toxicity index.²⁹ We observed that all compounds are non-cytotoxic to uninfected THP1 macrophage cells at the maximum tested concentration of 10µg/mL. However, standard antileishmanial agents, Amphotericin B and pentamidine are potently cytotoxic to the intra-macrophage amastigote while HDAC8-selective PCI-34051 is still inactive (Table 2). The 3HP compound 4a is moderately active in similar manner to its effect on the promastigote and axenic amastigote stages. Other exceptions in the 3HP series are 3a, 5a and 13a which display moderate to good cytotoxic activities, despite their inactivity against the promastigote and axenic amastigote stages (Supplemental Info Table S1). The target(s) responsible for the moderate activity of these 3HP compounds is unknown at the moment since they are inactive against the HDAC isoforms tested. Except for 6b and 8b, which are inactive, all 3HPT HDACi have moderate to strong cytotoxic activities. The potency of compounds 3b, 4b, 10b and HDAC6-selective Tubstatin A was enhanced by 5- to 25-fold relative to their effects on the axenic amastigote (Table 2). Interestingly, SAHA display more than a 50-fold potency enhancement relative to its effect on the axenic amastigotes. The improved potency of these HDACi (comparing axenic amastigote and intramacrophage amastigote stage, Tables 1 and 2) could be due to drug-induced secondary effects on the macrophage. HDACi are known to alter macrophage phenotype and function through perturbation of cytokine production,^30,31 which might negatively impact the viability of the intra-macrophage amastigote stage. Alternatively, upregulation of reactive oxygen species (ROS) production is another mechanism of HDACi-induced cytotoxicity.³²

Table 2.

Antileishmanial Activities (µg/mL) against the Intra-macrophage Amastigote Stage of L. donovani.

Compound	Cytotoxicity (THP1) IC50 (µg/ml)*	Macrophage Amastigote IC50 (µg/ml)*
1b	>10	6.1
2b	>10	5.8
3b	>10	0.6
4b	>10	0.9
5b	>10	3.9
6b	>10	>10
7b	>10	3.7
8b	>10	>10
9b	>10	2.1
10b	>10	0.3
11b	>10	2.6
12b	>10	2.7
13b	>10	1.5
SAHA	>10	0.5
Tubastatin A	>10	1.9
PCI-34051	>10	>10
Pentamidine	>10	0.4
Amphotericine B	>10	0.2

^*Average ± SD of triplicate observations.

We showed in this contribution that a set of 3HPT-derived HDACi with HDAC6 and HDAC8 inhibitory activities, but lacking HDAC1 inhibitory activity, are cytotoxic to extracellular and intracellular stages of L. donovani. We subsequently parsed out the contribution of inhibition of HDAC6-like and HDAC8-like activities to the overall antileishmanial activities using Tubastatin A and PCI-34051, standard HDAC6 and HDAC8 selective HDACi. We showed that Tubastatin A and 3HPT compounds are cytotoxic to both the extracellular and intracellular stages of L. donovani while PCI-34051 is nontoxic. We inferred from our observations in this study that HDAC6-like activity, to a larger extent, is essential for the viability of the amastigote and promastigote stages of L. donovani. Targeting protozoan HDAC6 with isoform selective inhibitors may then offer a more focused therapeutic strategy for Leishmaniasis.