The novel structure make LDM effectively remove CD123+ AML stem cells in combination with interleukin 3

Abstract

CD123 became a therapeutic target for acute myelocytic leukemia(AML) because of its overexpression only on AML stem cells. It is α subunit of interleukin-3 (multi-CSF, IL3) receptor. Lidamycin(LDM) is a novel antibiotic composed of an apoprotein (LDP) and a chromophore (AE). We cloned, expressed and isolated IL3LDP fusion protein first then assembled with AE in vitro. We found that131/132 amino acids of IL3 were the key factors for IL3 fusion protein stability and I131L/F132L mutation effectively improved the IL3 fusion protein stability. The toxicity of IL3LDM to CD123+ tumor cells was 2–10 times compared to LDM alone and 10000 times compared to ADR. Meanwhile, IL3LDM impaired the colony-forming ability of CD123+ stem-like cells but not to CD123 negative normal cord blood cells. Three drug delivery methods in vivo were adopted: prophylactic treatment and single/multiple-dosing administration. The tumor-free survival extended to 120 d and cancer cell invasion significantly decreased after IL3LDM continuous multiple treated. Moreover, IL3LDM had been shown to modulate apoptosis by arrested cell cycle in G2/M phase. Therefore, IL3LDM is expected to be a new drug for leukemia target therapy.

Abbreviations

AML

Acute myeloid leukemia

HSCs

Hematopoietic stem cells

IL3

interleukin 3

LDM

lidamycin

LSCs

leukemia stem cells.

Introduction

Experimental evidence has shown that acute myeloid leukemia (AML) is the result of haematopoietic stem cells (HSCs) malignant transformed following the accumulation of multiple stepwise genetic and epigenetic changes.¹ These transformed cells were considered as leukemia stem cells (LSCs) which were the first confirmed cancer stem cells (Casks). LSCs accounted for only 0.5% of total cells and are clonogenic with the potential for repetitive colony formation in vitro and the ability to establish leukemia when transplanted into immunodeficient mice.² It is the effective tumor destroyed by eliminating LSCs that changed leukemia treatment strategies.

CD123 is the only main positive marker on the LSCs but not on HSCs.³ CD123 is the α subunits of the interleukin-3 receptor(IL-3R) which is essential for IL3 binding.⁴ At present, the method to remove LSCs is mainly via IL3 or CD123, such as DT388IL3 fusion protein,⁵ IL3 antibody,⁶CD3×CD123 bispecific antibody,⁷ anti-CD123 CAR-T⁸ and so on. Therefore CD123 has become a promising target of anti-leukemia therapy with sparing normal progenitors. However, in the scenario of AML immunotherapy, Phase I clinical trials targeting CD123 by mAbs and immunotoxins (Clinical Trials.gov ID NCT 004401739, 01632852 and NCT 00397579)^9-10 have registered only minor clinical responses, suggesting the need to develop more powerful strategies.

Lidamycin(LDM, also known as C-1027) is worthy of paying attention as part of immunoconjugate because of its strong antineoplastic activity and special structure. As one of the most potent antitumoral agents ever discovered, LDM is currently undergoing phase II clinical trial in China.¹¹ After combined with anti-CD20Fab¹² or anti-type IV collagenase monoclonal antibody,¹³ AE exhibited marked inhibition on a panel of transplantable tumors in mice, which include leukemia L1210, P388, ascites hepatoma H22, sarcoma 180 and melanoma Harding-Passey.

Also, and importantly, LDM consist of 2 independent parts: an apoprotein moiety(LDP) and a non-protein chromophore extractable (AE) which could be reassembled and depart freely.¹⁴ The characteristic of its structure makes LDM extremely useful for gene project medicaments.

In this study, we produced IL3LDM fusion protein at first time and improved its stability with I131L/F132L mutation. In the following experiments, IL3LDM has been proven effective on CD123+ AML cells in vivo and vitro. LDM had a chance to participate more in killing CD123+ leukemia stem cells while spare the CD123 negative normal stem cells by IL3 delivery.

Result

Expression and characterization of IL3LDP and its mutant

The pET28a plasmid encoded IL3 followed by LDP and G4S linker(Fig. 1A). After subjected to Ni+ affinity chromatograpy, SDS-PAGE showed that IL3LDP fusion protein was observed at 27000Da which consistent with the expected molecular weight. And contrary to expectations, we found that a small molecular weight protein about 12000Da also reacted with anti-his6 antibody(Fig. 1B). We isolated and identified this ingredient by Protein spectrum. The result showed that the lower molecular weight ingredient was the latter part of the IL3LDP broken from 131 and 132 site of IL3(Fig. 1C). We designed a new primer for IL3LDM mutant to replace 131 and 132 bases of IL3 by 2 random bases. 78 clones were obtained by antibiotics screening and we picked out 3 more stable transformed clones(No.26, 47, 72) (Fig. 1D). The protein yield of original and 3 seed clones were 0.895, 1.206, 0.9815 and 0.8195mg/l, respectively. Protein purities were 54%,55%,92% and 59%,respectively. The sequencing of transforming clones were A131N in No26, A131L and I132L in No47,A131N and I132Y in No72(see Table 1). Purified proteins exhibit perfect binding activity with CD123(IL3Rα) on TF1 or MO7e cell line. Other than that, the protein produced by clone 26 had a moderately reduced affinity with TF1 and increasing affinity with MO7e cell line(Fig. 1 E). So the protein produced by No.47 clone have most stable performance and dramatically purity which is necessary for its antitumor activity research.

Figure 1.

Characterization of IL3LDP and its mutant. (A) The pattern of the recombinant plasmid; (B) The SDS-PAGE(12%) and western of IL3LDP(Line 1:bacteria periplasm lysates;Line 2:dialysate;Line 3:protein purification;Line 4:effluent liquid.); (C) Identification of low molecular protein broken form IL3LDM by Protein spectrum; (D) SDS-PAGE(12%) and western of IL3LDP mutants (a:No.26 clone;b:No.47 clone;c:No.72 clone. Line 1:bacteria periplasm lysates;Line 2: protein purification;Line 3: effluent liquid. Molecular weight standards were run on each gel left.); (E) The binding ability of purified fusion protein from different clones with IL3 receptor. (a:TF1 cell line;b:MO7e cell line. From left to right were blank, original clone, No. Twenty-six clone, No.47 clone and No.72 clone in cell lines.)

Table 1.

Parameters of fusion protein from different clones

Clone No.	BCA(ug/ml)	total output(mg/L)	IL3 ELISA(ug/ml)	purity	sequencing result
original	179	0.895	97	54%	A(131) and I(132)
26	1724	1.206	952	55%	A 131 N
47	981.5	0.9815	902	92%	A 131 L and I 132 L
72	1639	0.8195	968	59%	A 131 N and I 132 Y

Fusion protein binding affinities assay

Immune phenotypes were assayed in 6 leukemia cell lines by flow cytometer. It showed that the interleukin-3 receptor α chain(IL3Rα or CD123) was strongly expressed in KG1a, M07e, TF1 cell lines (80+5% positive) and weak positive in HL60 cells(7.8% positive). Conversely, Raji and K562 cell lines are CD123 negative(Fig. 2A). Unlike expected, most of AML primary cells were CD123 positive not just CD34+CD38− stem cells(Fig. 2E). IL3LDP(No.47) exhibited excellent IL3 receptor affinity on MO7e cell line(Fig. 2B) or clinic specimens(Fig. 2F). At the same time, we also found that the fluorescence intensity continue increased in dose-dependent manner even all cells have been combined with target protein(Fig. 2B). We also observed that IL3LDP protein bind to receptor molecules on the surfaces of cells under fluorescence microscope (Fig. 2C) and confocal microscopic system (Fig. 2D).

Figure 2.

Binding affinity of IL3LDP with cell lines and AML samples. (A) The expression level of CD123 in different leukemia cell lines and binding ability of IL3LDP(No.47) with these cells; (B) The binding ability of IL3LDP (No.47) by concentration decline with MO7e cells(blank, 20 ug,15 ug,10 ug,5 ug,2.5ug,2 ug, 1 ug, 0.1 ug/1×10⁶ cells/100ul); (C) The binding ability of IL3LDP (No.47) fusion protein with MO7e cells is observed under fluorescence microscope; (D) The binding ability of IL3LDP(No.47) fusion protein with MO7e cells is observed under confocal microscopic system; (E) The expression level of CD123 in CD34+/CD38−, CD34−/CD38−,CD34+/CD38+ and CD34+/CD38+ AML specimens detected by FACS; (F) The binding ability of IL3LDP(No.47) fusion protein with IL3 receptor on AML samples(blank, 10ng and1ug/1 × 10⁶ cells/100 ul).

Cell cytotoxicity assay

Four leukemia cell lines were used as the IL3 receptor-positive or negative cell line for measurements of inhibitory concentration

(IC₅₀s). The IC₅₀S of IL3LDM were 50.33,25.50,13.52 and 131.6 pM for MO7e,KG1a,HL60,Raji respectively. The IC₅₀S of LDM were 116,245.7,12.49 and 142.4 pM, respectively(see Table 1, Fig. 3A). The cell cytotoxicity of IL3LDM is 2–10 times as much as LDM in CD123+ cell lines and the difference is not significant in CD123− cell line. In human leukemic cells,HL60 cell line is the most sensitive cell to LDM(Fig. 3A). When we shorten the incubation time to 1h, the IC₅₀S of IL3LDM could be more than 100 times than LDM and drug effection reduced, too(see Table 2). At the same time, antitumor activity of IL3LDM is almost 7000 times as much as ADR whose IC₅₀S on cell lines are all above 100000 pM (Fig. 3B).

Figure 3.

Cell cytotoxin of IL3LDM, LDM and ADR in different cells and clinic samples. (A) IC₅₀s of IL3LDM, LDM and ADR for different cell lines; (B). Annumerical illustration for the above mentioned results; (C) Statistical chart of clone numbers with different treatment; (D) A comparative analysis of morphology change in survival clones; (E) The effects of IL3LDM to normal cord blood by AML-CFC assays.

Table 2.

IC₅₀s of IL3LDM, LDM and ADR co-incubated with cells for 4 hours (or 1 hours)

(pm)	IL3LDM	LDM	ADR
M07e	50.33 + 3.77(0.9376+0.125)	116.0 + 5.65(211.1+7.71)	351128 + 10.7
KG1a	25.50 + 1.13	245.7 + 6.67	347850 + 9.59
HL60	13.52 + 0.73(546.9+6.48)	12.49 + 0.49(1186 + 8.89)	111370 + 13.7
Raji	131.5 + 3.57	155.0 + 6.62	474106 + 11.7

AML-CFC test was carried on in order to verify the characters of CD123+ leukemia cells after co-cultured with IL3LDM or LDM. We chose CD123 high expression AML samples basing on our previous knowledge of their grow ability in NOD/SCID mice(see Table 3). Cytotoxicity of IL3LDM on clone forming was 3–5 times of LDM with the clone numbers(treated with IL3LDM(range,30–100AML-CFCs/10⁵cells) or LDM (range,70–200 AML-CFCs/10⁵ cells) )significantly declined compared with control(range,250–500AML-CFCs/10⁵ cells) (Fig. 3C). We also observed that the survival clones are small and disable after IL3LDM treated under different amplification (4×,10×,40×) in the same vision (Fig. 3D). Meanwhile,IL3LDM was nonpoisonous to normal cord blood cells which still have power to grow and differentiate. There were significant differences between each group (P < 0.05) (Fig. 3E).

Table 3.

AML Patients characteristics

sample	sex	diagnosis	CD123+(%)
1	male	AML	95%
2	male	AML	93.1%
3	male	AML	69.4%
4	female	AML	99.5%
5	Normal cord blood		0

NOD/SCID assay

AML cells (1×10⁷) from one patient sample with highest CD123 expression were transplanted into cohorts of 5-8 NOD/SCID mice after incubated with IL3LDM or LDM 2h(we called pretreatment group) or not(we called treatment group). AML patient cells labled by anti-human CD45 can be detected in the untreated mice bone marrow and tail blood after 6 weeks (Fig. 4A).

Figure 4.

NOD/SCID assay for IL3LDM. (A) Engraftment of AML patient samples was detected in the BM(above) and tail blood(below) at 8 weeks; (B) Percentage of PEcy5-CD45 labeled human leukemic cells in mice BM with different treatment; (C) Survival curve of mice with different treatment before killed at 12 weeks; (D) Pathological section of various organs from mice with different treated groups(Control and IL3LDP treated group: a large number of leukemia cell invasion to spleen and liver, obvious lung inflammation; pre-treated with LDM or IL3LDM: pathological analysis is normal; pos-treatment with LDM or IL3LDM: a large number of leukemia cells invasion in spleen but not in liver; Intensive pos-treatment with IL3LDM: pathological analysis is normal. In addition, haematopoietic cells displayed proliferation activity in spleen from IL3LDM treated group).

As compared with these controls, the mean percentage reduction of tumor cells in BM is 93%(range,78 to 96%) at week 8 and 93%(range, 90 to 97%) at week 12 in IL3LDM-pretreated mice respectively. At the same time, it is 81%(range,66 to 87%) and 84%(range, 53 to 93%) in LDM-pretreated mice, respectively (Fig. 4B).

Treatment group were treated by IL3LDM or LDM 4pmol/g/day×5 i.p after injected with AML cells. The proportion of AML patient cells in mice BM showed 35%(range, 31 to 45%) reduction in IL3LDM group and 46%(range, 63 to 87%) in LDM group at week 8 respectively. This effection maintained so that it was 65%(range,57 to71%) in IL3LDM and 54%(range, 37 to 79%) in LDM at week 12,respectively(Fig. 4B). Both two common ways can effectively control the disease development within 8 weeks after transplantation but leukemia cells regrowth after 12 weeks. Preventive treatment have better effect than treatment after transplantation(p < 0.5) on this point but there is a certain distance from the complete remission. However, differences between IL3LDM and LDM treatment is not significant. Patient cells were detectable easily in bone marrow with IL3LDP treated which agree with in vitro(Fig. 4B).

In order to consolidate curative effect, cohorts of 5 mice were treated more intensively with 3 5-day cycles every 20 dIt is observed that 60% leukemic cells (range, 40 to70%) eliminated in 8 weeks and 89%(range, 84 to 92%) in 16 weeks(Fig. 4B).

We also investigated the mice survival period and found that control group all died before 12 weeks. At the same time, the mice treated with IL3LDM 3 cycles can live until 16 weeks, although accidental death was observed during 10–16 weeks. Meanwhile, mice were all died before 8 weeks undergoing intensive LDM treatment(Fig. 4C).

Pathological section of various organs were made in order to determine whether IL3LDM could effective inhibit leukocyte invasion. The pathological reports are normal in IL3LDM or LDM preventive treatment group, by contrast, the curative effect is very poor and invasion was found in spleen but not in liver with drugs treatment after transplantation. Although no significantly different between IL3LDM or LDM treatment in each group, mice with continue IL3LDM treated showed better result whose pathological reports are all normal. Meanwhile, what is worth mentioning is that haematopoietic cells displayed proliferation activity treated with IL3LDM(Fig. 4D).

Machanism assay

Inserting the DNA minor groove and cleaving DNA is the main mechanisms of LDM(ZHEN YONGSU, 1998). The results of flow cytometric show that the number of viable cells treated by IL3LDM was smaller (73.3% at 24 h, 47.3% at 48 h and 42% at 72 h respectively) than LDM treated (65.4% at 24 h, 53.6% at 48 h and 48.6% at 72 h respectively). The total number of apoptotic cells (in LR and UR) and dead cells(in UL) remarkably increased after IL3LDM treatment (Fig. 5A).

Figure 5.

The killing mechanism of IL3LDM. (A) Analysis of apoptosis by flow cytometry at different times in MO7e cells with IL3LDM and LDM treated; (B) Analysis of cell cycle by flow cytometry at different time in MO7e cells with IL3LDM and LDM treated.

Cell cycle of MO7e cells was assay and IL3LDM arrested cell cycle in S phase, especially in G2/M phase. The proportion in S population was 47.84% at 24 h,46.42% at 48 h and 25.14% at 72,respectively. Whereas, it is 55.59%,66% and 23.7% in LDM groups. G2/M population with IL3LDM treatment was 3.56% at 24 h,23.38% at 48 h and 70.38% at 72 h. Whereas, it is 3.07%,7.81% and 50.45% in LDM groups (Fig. 5B). IL3LDM killed cells also by the way of cell cycle arrest and apoptosis just like LDM done, but made more effectual effort.

Discussion

Acute myeloid leukemia(AML) occurs in people of all ages, only 30–40% of AML patients younger than 60 y remain long-term survivors, furthermore, it represent only 10–15% of patients older than 60 y of age and <10% of patients with relapsed AML.¹⁵ Treatment outcome of AML, particularly in older patients, has not improved in the past 20 y

In the few past years, LSCs are thought to be the origin of the leukemia and derived from normal haematopoietic stem cells. High proportion of CD34⁺CD38⁻ LSC in AML diagnosis is reported to be associated with leukemia relapse and poor survival rate after chemotherapy treatment.¹⁶ The specific phenotypes of the AML-LSC are: CD34+, CD38−,CD71−, HLA-DR−, CD90−, CD117−, and CD123+¹⁷. The very first study (Frankel et al.) targeted AML-LSC with diphtheria-toxin interleukin 3-fusion protein(DT388IL3) showed such well results that CD123 was thought to be a promising target of anti-AML therapy.⁵ In the next few years, an neutralizing monoclonal antibody(mAb,7G3) targeting of CD123 produced by LiQ Jin et al. which effectively eliminates AML-LSC through blocking the homing effect of LSC as well as the activation of antibody-dependent cellular cytotoxicity (ADCC).⁶ Up till now both DT388IL3 and 7G3 have been used in phase I or II clinic trials.¹⁸

However, many problems still exist in the treatment described above. The response rate to DT388IL3 is low, only 2 cases of complete remissions and several partial remissions after DT388IL3 treatment.¹⁹ CD123 antibody–antigen complexes also have low internalization efficiency.²⁰ CD123xCD16 bispecific antibody reported that recruit NK cells for tumor cell killing, however, when PBMC was used in their studies, the required E/T ratio was above 20 which has had above average levels of clinical tolerance.²¹

The same problems persist in CD123xCD3 bispecific scFv designed by Kuo et al.⁷ and CD123XCD33 bispecific scFv.²²

Anti-CD123 CAR(chimeric antigen receptors) cells (Tettamanti S et al.) were able to strongly kill CD123⁺ cell lines which are more effectively than antibody-based approaches. But bone marrow suppression often go together.²³ The discovery and achievement above in recent years play an increasing role in cancer treatment and pave the way to develop a novel therapy approach for AML treatment by CD123 targeting, however they showing clinical efficacy as single agent remains limited.

Antibody-drug conjugates(ADCs)can enhance therapeutic potential of target drugs with conjugate to small molecule drugs and was still considered to be one of the most important anti-tumor techniques in future.

Lidamycin (LDM, originally named C-1027) as a potent enediyne-containing antitumor antibiotic, exhibits extremely potent cytotoxicity toward various human cancer cells (whether sensitive or multidrug-resistant cells) in vitro and vivo. LDM is about 1000-fold more potent than mitomycin C and adriamycin in human hepatoma cells and is an extremely potent antitumor agent that causes double-stranded DNA cleavages in every stages of cell cycle.²⁴ It was important that low concentrations LDM could also down-regulated embryonic stem (ES) cell-like genes Oct4, Sox2 and c-Myc.²⁵ LSCs are not sensitive to conventional chemotherapy drugs because they usually stay in the resting period. That suggests that LDM has better effection on LSCs.

Designing an ADCs is more complex than a simple meccano game, requiring thoughtful combination of antibody, linker, and drugs in the context of a target and a defined cancer indication. LDM is a unique small molecule-protein complex composed of a highly reactive enediyne chromophore, which upon binding reacts with its target molecule DNA through radical-mediated hydrogen abstraction and an apoprotein that encapsulates the chromophore serving as its carrier to reach DNA.²⁴ Many evidences indicate that LDM exhibit good activity after coupled with various molecules. We constructed the IL3(G₄S)LDP fusion protein expression plasmid with linker peptide size within 10 amino acid²⁶ and purified protein by anti-his6 Nickel ion affinity chromatography.

In our study,IL3LDM fusion protein was not stable and broke from the end of IL3 which has also happened in DT388IL3. Arthur E. Frankel et al. suggested that this process probably occurs in bacteria since the fragments were observed in cell pelletes prior to extraction.²⁶ Mutagenesis studies shows by Arthur E. Frankel et al. that two variants including deletion of C-terminal 8 amino acid residues or K116W in IL-3 resulted in higher-affinity interactions with IL-3R.²⁷ Although this conclusion has a significant guiding on IL3 protein stability, we are not sure the key amino acids. In this research, we identified the fall off part of IL3LDP fusion protein by protein mass spectrometry and compared its molecular weight with the theory value. Finally, we found that hydrolysis occurs at 131–132 sites in IL3. Meanwhile, LDM shows some aminopeptidase activity.^28-29 So we hold the opinion that the last 3 amino acids of IL3 have more effect on protein stability than binding ability. We replaced it with 3 random amino acids and screened 78 positive clones. We observed that protein hydrolysis significantly reduced in clone 26\47\72 by SDS-PAGE and western. It prompt that mutation of 131 to 133 bases in IL3 maybe a better way for IL3 fusion protein stability.

In subsequent experiments, we found that the protein produced by clone 47 had better affinity with a variety CD123+ cell lines. It is same as clinic samples after FACS and confocal analysis. Reaction with excess IL3LDM fusion protein could increase the fluorescence intensity on CD123+ cell surface even the cell positive rate was already 100%. This interesting phenomenon maybe suggested that more than one CD123 moleculars on one cell surface which was favorable to IL3 targeted therapy.

LDM showed extreme cytotoxicity toward various cultured cancer cells including leukemia L1210 and multiple myeloma U266.³⁰ We also confirmed this effection of IL3LDM in several AML cell lines. IL3LDM was more efficiently on killing CD123⁺ leukemia cells than CD123⁻ cells and its effection is 2–10 times (MO7e,KG1a) compared to LDM alone. The IL3LDM effective on MO7e cell line was reduced because of the GM-CSF competitive binding with IL3 receptors which was necessary for M07e culture. GM-CSF and IL3 shared the same receptors on cell surface. However, the effects on negative cells are not obvious. Better cancer-killing effects with less drug toxicity gained with IL3 guide.

Cytokine and receptor binding usually complete within half an hour to 1 hour while the process of cytotoxic drugs enter cells needs 2 hours at least. Based on the time difference between these 2 processes, the best effect of IL3LDM target therapy can be achieved within 1.5 hours. When we cut the incubation time to 1 hour, the cytotoxicity of IL3LDM is 100 times as much as LDM. So the different incubation time will cause different toxicity of IL3LDM and LDM.

CFC test is the “gold standard” to analyze stem cells. In this study, we take advantage of this experimental method to test the effects of IL3LDM on properties of CD123+ cells in vitro and found that IL3LDM treatment could obviously reduced the colony-forming efficiency of CD123+ cells. The form of residual clones were completely different from normal clones. So, IL3LDM could diminish CD123+ leukemia stem cells self-renewal ability and differentiation ability. Although CD123 as the marker of LSCs needed to be further confirmed, our experiments have shown that clearance of CD123 + cells does influence the formation of the malignant clone and we believe that it was really important to leukemia treatment. Meanwhile, CD123 didn't expressed on the surface of normal umbilical cord blood cells and we find that IL3LDM has little effect on normal haematopoietic cells. This can be effectively avoided a serious side effects.

IL3 or GM-CSF is necessary for AML cell line culture and it's transplantation must be done in mice bearing human IL-3 and GM-CSF genes (NOG IL-3/GM-Tg).³¹ To observe the activity of IL3LDM in vivo, we have no choice but developed another AML experimental mice model by many AML patient cells transplantion. Fortunately, several samples were transplanted successfully. In vivo, both IL3LDM and LDM are good at control the growth of transplanted tumor cells. The proportion of tumor cells in mice BM decreased which is same as prophylaxis and treated groups. No significant differences between IL3LDM and LDM in 8 weeks, however, IL3LDM fusion protein has a better effection after 8 weeks. At the same time, the effection of prophylaxis seems to be better than treatment. It was due to the more expand interact between tumor cell and drug in vitro. However, disease has an recurrence tendency at 8 weeks.

We believe that once treatment is impossible to completely remove leukemia stem cells, so the escaping cells cause disease recurrence. Pathological results showed that a large number leukemia cells invision in spleen, even proportion of human CD45+ cells in BM declined in prophylaxis or single-dosing administration groups. So we adopted multiple-dosing administration with IL3LDM and best result was obtained without tumor cells multiple organ invasion at 120 d Our result show, continuous administration with IL3LDM after transplantation prevent the disease relapse in a long period.

It is worth mentioning that the haematopoietic cells in spleen displayed proliferation activity in IL3LDM treated groups. We hypothesized that IL3LDP in blood could promote normal cells hematopoiesis after losing activity group(AE). This result is in agreement with the results obtained by IL3LDP treated alone in vitro(complementary1). And that's another evidence of the targeting of IL3 fusion protein.

We considered that the only side effect of IL3LDM is the damage to other normal CD123+ cells. IL3 is almost undetectable in the blood at very low levels, however, it is highly expressed on the 4 AML clinical specimens as we detected. The data from http://clinicaltrials.gov of DT388IL3 which had entered the phase II clinical trials has partly proved that IL3 fusion protein could be tolerate during clinical test procedures. The dose we used in IL3LDM treatment could be halved or quartered. Compared to worry about the side effects of IL3 fusion protein, we should be more concerned about how to improve its utilization in the body so that IL3LDM can work well in vivo as good as in vitro.

The percentage of CD34⁺CD38^low/−CD123⁺ leukemic cells at diagnosis was significantly correlated with response to treatment and survival.³² Another reported that AML-CFC cytotoxicity did not have progenitor killing activity in the long-term in vitro or in vivo assays. It is tempting to speculate that elimination of progenitors detected after 6 or more weeks in culture or in mice is more clinically relevant than toxicity detected against leukemic blasts or direct colony-forming cells.³³ More attention should be paid to the detection of progenitors or CD34⁺CD38^low/−CD123⁺ leukemic cells after treatment. Positive rate of cell surface CD123 directly affect the killing efficiency. Acute myeloid leukemia (AML) is a clonal disorder of hematopoiesis characterized. Several aberrations can coexist in subclones of acute myeloid leukemia, making it is impossible that the disease can be inhibited by a single agent or target. Cytosine arabinoside enhanced the efficacy of diphtheria-growth factor fusion proteins in an immunodeficient mouse model of human acute myeloid leukemia.³⁴ So we also want to improve the efficacy of IL3LDM with combination to other chemotherapeutic drugs.

Although IL3LDM shows excellent killing activity of CD123+ cells, we can still find the tumor cells in bone marrow of NOD/SCID mice. Whether CD123 is the only marker for LSCs or not is need to be proved. This is the challenging for all targeted therapy. We cultured 2 groups cells separated into CD123+/− from AML cell lines KG1a and found they grow into mix cell populations again(completment2). It is reported that earlier normal myeloid progenitors (LTC-IC) are less likely to require IL3 and express IL3 receptors,³⁵ so CD123 is important for stem cells yet not be the only accurately distinguished marker for the leukemia stem cells. This is one of the main factors to restrict the use of IL3 fusion protein.

AE is the activity group which is the small molecular and easily enter the tumor cells, while LDP act as protective protein does not need to enter the tumor cells. However, neither the apoprotein alone nor the chromophore alone is effective. LDM shows extremely potent cytotoxicity toward various cultured cancer cells even multidrug resistant cells. LDM rapidly and strongly inhibits DNA and RNA synthesis in hepatoma cells, subsequently blocks the cells at S and G2/M phase.¹¹ There was report of markedly cellular DNA strand scission in HL60 leukemia cell lines caused by a short-term LDM treatment (2 hours) at low concentration (5 nM).³⁶ Additionally, we have ascertained that IL3LDM had the same mechanism as LDM which is the indirect evidence of high efficiency of IL3 carrying the drug arrive at target cells. The delay in apoptosis to 48–60 h corresponds with the time required for DT388 toxin internalization, translocation and EF2 inactivation,³⁷ meanwhile half of the AML cell lines and samples were resistant to DT388IL3.³⁸ It is also reported that high-level CD123 expression on AML cells correlates with resistance to apoptosis⁴ which maybe one of reasons that CD123 positive cells were not sensitive to the general chemotherapy. Considering our results above, LDM holds the advantage that it can quickly enter the cells and inserted into DNA minor groove, regardless of the cell types and the period of cell cycle which is important to multi-drug resistant tumor stem cells even AML stem cells. Even more interestingly, LDM downregulated the genes related to stem cells including Oct4, Sox2 and Myc so that it inhibited mouse embryonic carcinoma (_EC) cell growth.²⁵ It preferred to weak cell self-renewal ability rather than cause cell apoptosis or death. Sensitivity to the stem cells or stem like cells made LDM more suitable for application in LSC-targeted therapy. Meanwhile, AML model by overexpressing the human mixed-lineage leukemia-AF9 (MLL-AF9) fusion gene in mouse haematopoietic cells that carry Yamanaka factors (Oct4, Sox2, Klf4 and c-Myc)under the control of doxycycline (Dox). Interestingly, most chimeric mice spontaneously developed the same type of AML.³⁹ This may be explained at least partly LDM has high sensitivity to LSCs. The human interleukin-3 (IL-3) receptor (IL-3R) is a heterodimeric structure composed of α and β(βc) subunit, both are required for signal transduction by the IL-3 cytokine.⁴⁰

QRT-PCR levels for both the IL-3R α and βc subunit expressions correlated significantly (r = 0.59 and 0.70,respectively) with AML-CFCs killed with DT388IL3.⁴¹ However, in this study the relationship was not investigated thoroughly. Perhaps we should screen appropriate patients by detecting the βc subunit expression level before given IL3LDM treatment to improve effection. Taken together, all the data support the notion that it is a approachable way to target CD123+ acute leukemia stem cells by the IL3-based therapy.

We're not sure yet whether leukemia is the result of clonal evolution or cancer stem cells ?We demonstrated, for the first time, that LDM was a novel “warhead drug” against cancer stem cells whose effection can be amplified by IL3 target system. To our knowledge, in-depth analysis in applicable persons maybe the necessary way for this individualized target treatment.

Conclusions

CD123(α subunits of IL3R) is the only main positive marker on the LSCs but not on HSCs. This discovery displayed a brilliant prospect of anti-LSCs medicine based on IL3. We produced IL3LDP fusion protein and identified the key amino acids which influences the protein stability at first time We observed that IL3LDP protein bind to receptor molecules on the surfaces of cells. In the following experiments, we proved that IL3LDM with I131L/F132L mutation has good inhibitive effect on CD123+ AML cells in vivo and vitro. Flow cytometry and pathological results showed that multiple- dosing administration of IL3LDM have best curative effect in the 3 ways (prophylaxis, single-dosing administration or multiple- dosing administration). IL3LDM killed cells also by the way of cell cycle arrest and apoptosis just like LDM done. LDM had a chance to participate more in killing CD123+ leukemia stem cells with IL3s help while spare the CD123 negative normal stem cells. IL3LDM may be an effective method to remove LSCs and plays an important role in leukemia treatment.

Material and Method

Plasmid construction

LDM was isolated from a culture filtrate globisporus

(Preserved by Institute of medicinal biotechnology Academy of Medical Sciences & Peking Union Medical College) as reported in the previous study[2]. An IL3 DNA cloned from human peripheral blood mononuclear cells.

Primers(Synthesis by invitrogene) were used to introduce MluI and SphI restriction sites at the ends of the IL3 and LDP gene so that these DNA were synthesized by PCR(Fig. 1A).

Primer for IL3

5′-GCGCACGCGT CGCTGCTCCCATGACCCAGACAACG

3′-CGCTGATCCGCCTCCACCAAA GATCGC GAGGCTCAA AGT

Primer for IL3 mutation

3′-GCGGGCGCTGATCCGCCTCCACCAAA NNNNNN GAGGCTC

Primer for LDP

5′-TTTGGTGGAGGCGGATCAGCGCCCGCCTTCTCCGTCAGT

3′-1GTGATGTGATCCGCCTCCACCGCCGAAGGTCAGAGCCAC

3′-2GCGCGCATGCTCAATGATGGTGATGGTGATGTGATCCGCCTCCACC

Expression and purification of IL3LDP and its mutant

Commercialization pET28a(clontech) joining the IL3LDP periplasm cavity expression guided by signal peptide was transformed into BL21. Colonies from 100 mm LB kanacillin plates were pooled and shaken in 1L fernbach flasks at 37°C. When the culture reached an OD650 of 0.6, 1 mM IPTG (Sigma) was added. Four hours later, the cells were harvested by centrifugation, resuspended in 25 ml lysis buffer (pH 7.5, 20%glucose,1 mM EDTA,200 mM NaCl,25 mm Tris,0.1mm PMSF), incubated 1 hour on ice and extracted the supernatant. Furthermore, the protein isolated in the periplasm can be purified by nickel affinity after dialysis 24 hours with 1 × PBS to removal salt. Finally, protein was enriched through ultrafiltration centrifugal tube(Millipore, 50 ml,10kD), measured the concentration or purity with BCA kit(Pierce) and IL3 ELISA kit (Neobioscience).

IL3LDM mutant were cloned and protein purified using the same method.

Characterization of IL3LDP and its mutant

Purified proteins were run on a 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and either stained with Coomassie Blue R-250 (Sigma Chemical) or transferred to nitrocellulose, blocked with 10% non-fat dry milk, washed with 1×PBS containing 0.05% Tween-20,reacted with a 1:1000 anti-his6 antibody (TIANGEN biotech), rewashed, incubated with sheep anti-rabbit IgG conjugate FITC antibody (Amyjet scientific Inc.) in 1:3000, washed again, development by DAB.

Non-target protein with small molecular weights was identified by mass spectrometry (Bioyong Tech, China) after separation from gel.

IL3LDP and its mutant receptor binding affinities

TF-1\MO7e\KG1a\HL60\Raji cells were maintained in the log phase of cell growth in RPMI1640 media with(TF-1\MO7e) or without(KG1a\HL60\Raji) 10ng/ml GM-CSF(Amoytop biotech, China). Peripheral blood samples from newly diagnosed AML patients after obtaining informed consent. Diagnosis and classification of AML were based on the criteria of the French-American-British group. Normal cells isolated from umbilical cord blood were utilized in our experiments and has been ethically approved and followed the regulation of Helsinki institute.

Cells incubated with PE-CD123 or its isotype (BD pharmingen,1×10⁶ cells/100ul/5ul antibody) at 4°C for 1 hour for identification of CD123 expression. Meanwhile, cells are incubated with purified proteins(1×10⁶ cells/100ul/1ug IL3LDP protein),anti-his6 antibody(TIANGEN biotech) and rabbit anti-rat IgG conjugate FITC(Amyjet scientific Inc.) in turn at 4°C for 2 hours, respectively. PBS washing required after every step. Immunophenotype was examined by flow cytometry(BD LSRII).

Binding capacity of fusion protein(No.47) by concentration decline(20 ug,15 ug,10 ug,5 ug,2.5 ug,2 ug,1 ug,0.1 ug/1 × 10⁶ cells/100ul) with MO7e cells was determined by the same method. Then we observed the cells under fluorescence microscope. We further observed cells under confocal laser with nuclei stained by DAPI (AppliChem, China). It is also done to detect the binding capacity of fusion protein(No.47) with AML cells in low dosage(10 ng/1 × 10⁶ cells/100 ul) and high dosage(1 ug/1 × 10⁶ cells/100 ul).

Assemble with AE

LDM (10 mg) with high potent activity were suspended in 5 ml methanol and whisked for 5 min at 4°C, and then the mixture was placed at 20 1C for 1 h. AE was obtained from the supernatant of reaction mixture by centrifugation at 16 000 r.p.m. for 20 min at 4 1C. The above procedure was repeated once to isolate AE completely.

For obtaining fusion protein IL3LDM, AE in methanol was added to IL3LDP/PBS (10 mmol/l, pH 7.0) with the molecular ratio of 5:1 and the volume ratio of 1:50, respectively, and then placed at room temperature for 12 h. Purified IL3LDM was prepared finally by separating from free AE by using a sephadex G-75 column.

Cell cytotoxicity assay

For cell counting kit-8 assay (Dojindo),1×10⁵ cells/ml were incubated in 100 μl RPMI 1640 with or without GM-CSF in a 96 well microplate. Make a well of only media to measure the background. Seven different concentrations of IL3LDM(from 0.5 to 35.5 nm,double dilution) or LDM(from 1.5 to 95nm, double dilution) were added to some column of the 96-well plate in 3–5 uL of media, maintained at 37°C, 5% CO₂ for 4 h. Media was then removed and add 100μl new media to each well. Meanwhile, different concentrations ADR (from 0.225 to 14.8 um,double dilution) was added to another cells without change new media. After continue culturing 72h,10 μl Cell Counting Kit-8 was added to each well. Place the plate in a CO₂ incubator for 3–4 hours to react. Measure the absorbance at 450 nm with a microplate reader.

To evaluate inhibition of colony formation as another measure of toxicity, 1–2×10⁵ AML and normal cord blood cells were incubated with IC₅₀ IL3LDM or LDM in 4 ml of RPMI 1640 medium in 6-well flat-bottomed costar plates at 37°C, 5%CO₂ for 1h. Cells were then harvested and mixed with 1 ml methylcellulose medium H4434 (StemCell Technologies, Canada). Poured into one well of 6-well flat-bottomed costar plates and sterile deionized water injection around. Plates were placed in humidified chambers at 37°C, 5%CO2 for 14 days, after which colonies containing greater than 20 cells were partition counted.

NOD/SCID assay

In preliminary experiments cohorts of 5–6 week-old NOD/SCID mice received different doses of LDM daily for 5 d i.p. Thus, the maximum tolerated dose was considered to be 4pmol/g of body weight, whereas at or below this dose,90% of treated mice survived with no apparent ill effects.

NOD/SCID Mice,4–5 weeks, received 300 cGy from a ¹³⁷Cs source 6–8 h before injection of AML patient cells with highest CD123 expression. AML cells incubated with or without 50 nm IL3LDM or LDM for 2 h before transplanted into mice via the tail vein. On the other hand, Mice treated with 2–4pmol/g of body weight IL3LDM or LDM daily of i.p. for 5 d at intervals after injection of untreated AML cells. To attempt to improve these results, cohorts of 5 mice were treated more intensively with 4pm/g 3 5-day cycles.

Half of the cells from mouse BM were incubated with a PE-Cy™5 mouse IgG1Κ isotype control (BD pharmingen) and the other half were incubated with PE-Cy™5 mouse antihuman CD45 antibody which is a human-specific pan-leukocyte marker (BD pharmingen) 30 minutes on ice. FACS analysis was done on a Becton Dickinson FACSort flow cytometer. Nonspecific binding of CD45 on mouse bone marrow cells is reliably <0.1%. Thus, a difference between the IgG1-negative control and the CD45 expression of the treated mice of >0.1% was regarded as evidence for engraftment of human cells.

Machanism assay

Cells were placed into 6-well plates to allow cell grow and cultured overnight, then were not treated (blank group) or treated with 50 nm LDM and IL3LDM for 1 hours. After changing new medium, cells were continue cultured and collected at 24, 48, 72 h, then washed 2 times with PBS. Detection was performed according to instructions of the cell apoptosis detection kit and cell cycle detection kit (Beyotime institute of Biotechnology). Changes of apoptosis and cell cycle were analyzed with flowjo7.6 software.

Statistics analysis

Results are indicated as the means ± SD. Treatment effects were compared using the T test and differences between the means were considered to be significant when P < 0.05.

Funding

This paper was supported by National Science and Technology major projects “major drug discovery” (2012ZX09102301) and China Nature Science foundation (81400176). Sources of material (LDM) are not available commercially.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.