Gene expression profiles associated with early relapse during first remission induction in canine multicentric high-grade B-cell lymphoma

Abstract

Although chemotherapy using CHOP-based protocol induces remission in most cases of canine multicentric high-grade B-cell lymphoma (mhBCL), some cases develop early relapse during the first induction protocol. In this study, we examined the gene expression profiles of canine mhBCL before chemotherapy and investigated their associations with early relapse during the first whole CHOP-based protocol. Twenty-five cases of mhBCL treated with CHOP-based protocol as first induction chemotherapy were included in this study. Sixteen cases completed the first whole CHOP-based protocol without relapse (S-group), and nine developed relapse during the chemotherapy (R-group). RNA-seq was performed on samples from neoplastic lymph nodes. Differentially expressed genes (DEGs) were extracted by the comparison of gene expression profiles between S- and R-groups, and the differences in the expression levels of these genes were validated by RT-qPCR. Extracted 179 DEGs included the genes related to chemokine CC motif ligand, T-cell receptor signaling pathway, and PD-L1 expression and PD-1 checkpoint pathway. We focused on chemokine CC motif ligand, and CCL4 was confirmed to be significantly downregulated in the R-group (P=0.039). We also focused on the genes related to T-cell signaling pathway, and CD3E (P=0.039), ITK (P=0.023), and LAT (P=0.023) genes were confirmed to be significantly upregulated in the R-group. The current results suggest that both changes in tumor cells and the interactions between tumor cells and immune cells are associated with the efficacy of the chemotherapy for first remission induction.

Lymphoma is one of the most common tumors in dogs, accounting for 83% of canine hematopoietic tumors [27]. Canine lymphomas are sub-classified according to the anatomical location of the lesions, immunophenotype, cell morphology, and histopathological examination [17, 34, 41, 47]. Multicentric high-grade B cell lymphoma (mhBCL) is the most common subtype in dogs [12, 34], and 75% of the cases with this subtype are classified as centroblastic type according to updated Kiel classification [25, 34, 41]. mhBCL generally corresponds to diffuse large B-cell lymphoma (DLBCL) according to WHO classification based on histopathological examination [2]. Chemotherapy is the treatment of choice for canine mhBCL to prolong overall survival [12, 18], and University of Wisconsin (UW)-Madison chemotherapy protocol (UW-25), a CHOP-based protocol comprising cyclophosphamide (CPA), hydroxydaunorubicin (doxorubicin, DXR), vincristine (VCR), and prednisolone, is one of the most commonly used protocols [12, 18]. Although 80–95% of the mhBCL cases treated by CHOP-based protocol can be led to remission, approximately 95% of the cases develop relapse due to chemoresistance [11, 18, 22, 32]. However, the remission duration is various among the cases, and part of the cases develop early relapse during the first induction of remission, which means early acquisition of chemoresistance.

In human medicine, alterations in the expression of various genes have been reported as the factors related to the development of resistance to chemotherapeutic drugs in various tumor types [5]. Genes associated with drug transportation, modulation of cell death, detoxification of anticancer agents and DNA repair are known as drug-resistant factors [5]. In veterinary medicine, overexpression of P-glycoprotein, a drug efflux pump encoded by ABCB1 gene, has been extensively investigated [4]. We previously reported that high expression of ABCB1 gene was observed in a subset of dogs with lymphoma that developed resistance to chemotherapeutic drugs [39]. Other drug-resistant factors, which were previously reported in human medicine, were not associated with chemoresistance in canine lymphoma in the same previous study [39]. Therefore, we investigated novel candidate genes associated with the acquisition of chemotherapy resistance by comparing gene expression profiles in canine mhBCL cases between before and after the acquisition of chemotherapy resistance [35, 38]. Other studies have shown that the expression of checkpoint molecule is increased when canine lymphoma cell lines acquired resistance to chemotherapeutic agents [20]. Additionally, it has been suggested that TGF-beta signaling pathway may be involved in resistance to doxorubicin in canine lymphoma cell lines [21]. However, there has been no study that investigated any molecules associated with early relapse during the first induction of remission in canine mhBCL cases.

The objective of this study was to investigate the differences in the gene expression profiles of canine mhBCL before chemotherapy between the cases that completed the first induction protocol and those that developed early relapse during the chemotherapy.

MATERIALS AND METHODS

Cases and samples

This study included 25 dogs presented to The University of Tokyo Veterinary Medical Center, Japan Small Animal Cancer Center, JASMINE Animal Referral Hospital, Bayside Animal Clinic, Japan Animal Referral Medical Center, Yamaguchi University Animal Medical Center, and Shikoku Veterinary Medical Center between 2018 and 2022 (Table 1). Written informed consent was obtained from the owners of all dogs before they were included into.

Table 1. Signalment of 25 dogs with multicentric high-grade B-cell lymphoma included in the present study.

Dog No.	Age (years)	Sex	Breed	Body weight (kg)	WHO stage	Group	RNA-seq	RT-qPCR	PARR at relapse
1	11.0	Female (spayed)	Mix	8.30	Va	S	○	○	NE
2	9.4	Male (castrated)	Golden Retriever	28.50	IVa	S	○	○	○
3	10.3	Male	Welch Corgi Pembroke	15.00	Va	S	○	○	○
4	12.2	Male (castrated)	Chihuahua	4.85	Vb	S	○	○	○
5	6.5	Male (castrated)	Mix	3.68	IVb	S	○	○	NE
6	9.2	Male (castrated)	Miniature Schnauzer	8.98	Va	S	○	○	○
7	5.8	Male	Mix	3.90	Va	S	○	○	NE
8	7.0	Male	Mix	7.75	Va	S	○	○	NE
9	11.3	Female (spayed)	Shiba	7.05	IVb	S	○	○	NE
10	11.9	Female (spayed)	Jack Russell Terrier	8.58	IVa	S	○	×	NE
11	6.5	Male	Pomeranian	2.80	Va	S	○	×	○
12	12.8	Male (castrated)	Poodle (Toy)	5.16	IVa	S	×	○	NE
13	11.3	Male	Shih Tzu	5.35	Vb	S	×	○	○
14	9.3	Male (castrated)	Shiba	13.08	Va	S	×	○	NE
15	13.2	Female (spayed)	Miniture Dachshund	7.70	Va	S	×	○	NE
16	10.7	Female (spayed)	Welch Corgi Pembroke	8.40	IVa	S	×	○	○
17	14.7	Male (castrated)	Shiba	11.20	IVa	R	○	○	NE
18	7.0	Female (spayed)	Poodle (Toy)	7.40	Vb	R	○	○	○
19	3.2	Male (castrated)	French Bulldog	13.80	Vb	R	○	○	○
20	11.3	Male	Shih Tzu	5.35	Vb	R	○	○	NE
21	12.9	Male (castrated)	Border Collie	19.00	Vb	R	○	○	○
22	10.3	Female (spayed)	Shih Tzu	5.54	Va	R	○	○	NE
23	12.5	Male (castrated)	Shiba	13.80	IIIa	R	○	○	○
24	11.9	Female (spayed)	Kai Inu	12.25	Vb	R	×	○	○
25	14.6	Male (castrated)	Chihuahua	3.92	IVb	R	×	○	○

S, S-group; R, R-group; PARR, PCR for antigen receptor gene rearrangements; NE, not evaluated.

The dogs underwent physical examination, complete blood cell count, thoracic and abdominal radiographs, abdominal ultrasound examination, and cytological examination of fine needle aspirate (FNA) samples of the swelling superficial lymph nodes (LNs). When abnormal findings were observed in the liver or spleen by ultrasonography, it was cytologically evaluated. The infiltration of tumor cells into peripheral blood was evaluated by peripheral blood smear, and bone marrow exam to evaluate marrow infiltration was not performed in any dogs. WHO clinical stage was determined based on the results of these examinations [33]. Cytological examination of FNA samples were performed by a veterinarian (H.T), and the morphological classification was confirmed as centroblastic type in all dogs based on the updated Kiel classification [17]. In addition, every FNA sample was cytologically confirmed that the majority of the cells collected were neoplastic lymphocytes. The immunophenotype of tumor cells was confirmed as B-cell origin using PCR for antigen receptor gene rearrangements (PARR) [19]. PARR analysis to detect the rearrangements of immunoglobulin heavy chain and T cell receptor gamma-chain genes was performed using GeneScan analysis, a capillary electrophoresis system that uses fluorescence-labeled primers and is capable of separating different PCR products that differ in size by 1 bp.

All dogs were treated with UW-25 protocol for initial induction of remission [18]. Cases that received any chemotherapy before sample collections were excluded. Veterinarians measured the size of the LNs with calipers at each admission, and response to treatment was evaluated based on response evaluation criteria for peripheral nodal lymphoma [40]. All dogs included in this study achieved complete remission (CR) by week 9 of the UW-25 protocol. Relapse was defined as the recurrence of the peripheral lymph node swelling and cytologic evidence of centroblastic type lymphoma, same as it was at the diagnosis. The dogs were classified as S-group if they maintained CR until week 25 of the UW-25, and the other dogs that developed relapse by week 25 of the UW-25 protocol were classified as R-group (Fig. 1).

Fig. 1.

Canine multicentric high-grade B-cell lymphoma cases included in the present study. The dogs were classified as S-group if they maintained complete remission until week 25 of the UW-25, and the other dogs that developed relapse by week 25 of the UW-25 protocol were classified as R-group.

Extractions of DNA and RNA

Tumor samples were obtained by FNA for the swelling LNs of the dogs enrolled. Those samples were immediately immersed in RNAlater (Sigma-Aldrich, St. Louis, MO, USA) and stored at −80°C. The specimens in RNAlater were washed three times with PBS and then total RNA and DNA were extracted from these samples using the Allprep RNA/DNA minikit (QIAGEN, Hilden, Germany) following the manufacturer’s protocol. DNA and RNA sample were immediately frozen until the following analyzes at −20°C and −80°C, respectively.

RNA sequencing (RNA-seq)

RNA integrity was examined with an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Sequencing libraries were prepared with the extracted total RNA using TruSeq Stranded mRNA Library Prep Kit for NeoPrep (Illumina, San Diego, CA, USA). RNA sequencing analysis (75 bp paired end) was performed on NovaSeq 6000 (Illumina) using the High Output Kit (Illumina).

Quality controls and adaptor trimmings of the obtained fastq files for each sample were performed using the Trim Galore software based on Fastp. Trimmed fastp data were mapped to canine genomes (CanFam3.1) by STAR V.2.7.3a,22 and transcript abundance was estimated using RSEM V.1.3.323 with gene transfer file for Ensembl (CanFam3.1.98, https://www.ensembl.org). Trimmed mean of M values normalization was conducted to adjust for abundance differences across the samples, and obtained gene count data were used to extract differential gene expressions with an EdgeR-based R package, TCC V.1.32.0. The differentially expressed genes (DEGs) were extracted in the comparisons between S- and R-groups by false discovery rate <0.2. To explore the biological mechanism and pathways associated with the extracted DEGs, gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID). P-values were adjusted by applying the Benjamini–Hochberg procedure with a threshold of P<0.05.

The datasets used and analyzed in the present study are available at the DDBJ Sequenced Read Archive repository with accession number DRA015976.

RT-qPCR

Reversed transcript was constructed using 10 ng of extracted total RNA and ReverTra Ace (TOYOBO, Osaka, Japan) following manufacture’s instruction. Primer pairs used for each gene are listed in Table 2. RT-qPCR was performed using TB Green Premix Ex Taq^TM II (Tli RNaseH Plus) (Takara Bio, Kusatsu, Japan) by Thermal Cycler Dice Real Time System TP800 (Takara Bio). TBP was used as an internal control gene based on a previous study [39], and the relative mRNA amounts of each gene were calculated using the standard curve analysis. All samples were examined in duplicate.

Table 2. Primer pairs used in RT-qPCR for the 13 genes.

Gene	Forward	Reverse	Product size (bp)
CD3E	TGGACGACAGTGGTTATTATGC	TAATACACCATCAGCAGCAACC	172
ITK	TCTGGTGGAAAAGTCTCCAAAT	GCTTGATACAGGGGTTGTTCTC	200
LAT	TCCTATGGGCCTGTTACCTCT	AGCCCTCGTTGTGATACTCCT	214
CD28	AACAGTGGCTCTTTGTGCCTA	CAGTTGTGTTGAGCAGGTGAA	213
ZAP70	AAGCATAGCACGGAGAAGACA	AGGTACTTCATCCCCATGGAC	242
PRKCQ	ACCCCAGAAAGAGAGGACAAA	GTCCCAGGGTGGGTACATAGT	219
PD-1	CTACTGCTGCTGCTGACCT	GATGGTGGCATACTCGGTC	216
CD8A	TCTGTCCTGAGCAACTCCATT	GGTGCCCAGATGTAGATTTCA	218
CTLA4	GGGAAATGGAACCCAGATT	AAACAAGCCCGAACTGACT	97
CCL3	CAAGCAGATTCCACGCAAGTT	TAATACCGGGCTTGGAGCAT	71
CCL4	CGTCCTTTCTCTCCTTGTGC	GAATCTTCCGCAGGGTGTAA	113
CCL5	GGTCTCCGCAGCTACCTTT	AAAGCAGCAGGGTGTGGT	100
ABCB1	ACTCGGGAGCAGAAGTTTGA	AATGAGACCCCGAAGATGTG	95
TBP	CTATTTCTTGGTGTGCATGAGG	CCTCGGCATTCAGTCTTTTC	96

Statistical analysis

Age, body weights and distribution of WHO clinical stage and substage were compared between the S- and R-groups by Fisher’s exact test. The relative amounts of each gene mRNA were statistically compared between the groups by Mann–Whitney U test. The statistical analyzes were conducted using R version 4.1.1 (R core team, Vienna, Austria), and a significance level of P<0.05 was used to indicate statistical significance.

RESULTS

Cases

Dose reductions of VCR, CPA, and DXR were performed in 16, 9, and 4 out of the 25 cases, respectively, due to the adverse events observed.

In all 25 cases, relapse occurred in the peripheral lymph nodes including the ones from which samples has been collected prior to chemotherapy.

PARR testing was performed at relapse in 13 cases, which confirmed that amplicon lengths of PCR products using initial and relapse samples were identical in each case (Table 1). In 12 cases, PARR testing was not conducted at relapse.

At first, RNA-seq was performed using the samples of 18 cases (Table 1 and Fig. 1). The most common breed was mixed breed. The median age of the dogs was 10.3 years (range, 3.2–14.7 years). The median body weight was 8.0 kg (range, 2.8–28.5 kg). There were 13 male dogs (5 intact, 8 castrated) and 5 female dogs (5 spayed). The 18 cases consisted of 11 cases in S-groups and 7 cases in R-groups. The cases included in the two groups were comparable regarding age (P=1), gender (P=0.24), and body weight (P=1), the distributions of WHO clinical stage (Stage V or others) (P=1), and clinical substage (P=0.33) (Table 3). Although all 11 cases in the S-group underwent drug reduction of any drugs due to adverse events, 4 of 7 cases in the R-group did not. The proportions of the cases with dose reductions of any drugs were significantly higher in S-group than in R-group (P=0.011) (Table 3).

Table 3. Comparisons of dogs with multicentric high-grade B-cell lymphoma included in RNA-sequencing between S- and R-groups.

Factor		S-group	R-group	P-value
Age				1
	>10.3 years	6	4
	≤10.3 years	5	3
Sex				0.242
	Male	4	1
	Male (castrated)	3	4
	Female	0	0
	Female (spayed)	3	2
Body weight				1
	>8 kg	5	4
	≤8 kg	6	3
WHO stage				1
	Not V	4	2
	V	7	5
WHO substage				0.332
	a	8	3
	b	3	4
Dose reduction
	Yes	11	3
	No	0	4	0.0114

Extraction of the DEGs and enrichment analysis

The RNA integrity number values exceeded 8.0 (range: 8.9–10) for all RNA-seq samples. RNA-seq analysis generated at least 15.96 million raw reads and >96.60% of mapping rates with paired reads for each sample (Supplementary Table 1). A comparison of gene expression profiles between S- and R-groups resulted in 179 DEGs, with 169 genes upregulated and 10 genes downregulated in the R-group against the S-group (Fig. 2 and Supplementary Table 2). The top five DEGs with corresponding gene names based on q-values were PRF1, GZMB, TG, PALMD, and DUSP4.

Fig. 2.

M-A plot of differentially expressed genes (DEGs) extracted by the comparisons between S- and R-groups. The 179 DEGs are shown in pink colors.

The GO analysis of the biological functions associated with the DEGs, including biological process, cellular component and molecular function, were shown in Table 4. These enriched terms were mainly associated with immune response including “immunological synapse”, “CCR chemokine receptor binding”, “chemokine activity”, “monocyte chemotaxis”, “neutrophil chemotaxis”, “lymphocyte chemotaxis”, “cellular response to interferon-1”, “cellular response to interferon-gamma”, “chemokine-mediated signaling pathway”, “immune response”, and “T cell co-stimulation”. In KEGG pathway analysis, four pathways, including “T cell receptor signaling pathway and PD-L1 expression” and “PD-1 checkpoint pathway in cancer”, were shown to be enriched (Table 5).

Table 4. Results of Gene Ontology analysis using the extracted differentially expressed genes by Database for Annotation, Visualization and Integrated Discovery.

Group	Term	Count	P-value
Cellular component	Immunological synapse	7	0.00016
Cellular component	External side of plasma membrane	13	0.0006
Cellular component	Extracellar space	25	0.0012
Cellular component	Extracellar region	17	0.022
Molecular function	CCR chemokine receptor binding	5	0.0064
Molecular function	Calcium-dependant protein binding	6	0.012
Molecular function	Chemokine activity	5	0.023
Biological process	Monocyte chemotaxis	6	0.0025
Biological process	Neutrophil chemotaxis	7	0.0025
Biological process	Lymphocyte chemotaxis	5	0.014
Biological process	Cellular response to interleukin-1	6	0.014
Biological process	Positive regulation of gtpase activity	8	0.023
Biological process	Cellular response to interferon-gamma	6	0.029
Biological process	Chemokine-mediated signaling pathway	5	0.04
Biological process	Immune response	9	0.048
Biological process	T cell co-stimulation	4	0.048
Biological process	T cell receptor signaling pathway	6	0.048
Biological process	Cellular response to tumor necrosis factor	6	0.048

Table 5. Results of pathway analysis using the extracted differentially expressed genes by Database for Annotation, Visualization and Integrated Discovery.

KEGG pathway	Count	P-value
T cell receptor signaling pathway	9	0.00052
Autoimmune thyroid disease	5	0.038
Rheumatoid arthritis	6	0.042
PD-L1 expression and PD-1 checkpoint pathway in cancer	6	0.042

KEGG pathway, Kyoto Encyclopedia of Genes and Genomes pathway.

Validation of differences in mRNA expression of DEGs between the groups

Of the RNA-seq cases taken, two from the S-group were excluded from RT-qPCR analysis due to insufficient RNA, and a further seven cases were added to this analysis; five from the S-group and two from the R-group (Table 1 and Fig. 1). Although 13 of the 14 cases in the S-group underwent drug reduction of any drugs due to adverse events, 4 of the 9 cases in the R-group did not. The proportions of the cases with dose reductions of any drugs tended to be higher in S-group than R-group (P=0.056).

At first, we focused on the extracted DEGs related to “chemokine activity” and “CCR chemokine receptor binding”, and CCL3, CCL4, and CCL5 genes analyzed using RT-qPCR were compared between 14 S-group cases and 9 R-group cases. CCL4 showed significantly higher expression levels in the R-group compared with S-group (P=0.038), while CCL3 and CCL5 did not show any significant difference between the groups (P=0.64 and 0.64, respectively) (Fig. 3A–C).

Fig. 3.

Comparison of the relative expression levels of the nine differentially expressed genes, CCL3 (A), CCL4 (B), CCL5 (C), ABCB1 (D), CD3E (E), ITK (F), LAT (G), CD28 (H), and ZAP70 (I), which were extracted between 14 cases in S-group and 9 cases in R-group. Statistically significant differences were observed in the relative amounts of mRNA of CD3E (P=0.039), ITK (P=0.023), LAT (P=0.023) and CCL4 (P=0.039) between S- and R-groups.

Then, we examined the mRNA expression levels of ABCB1 by RT-qPCR because this gene has been known to be associated with chemotherapy resistance and extracted as one of the DEGs in this study (Supplementary Table 2). There was no significant difference in the relative amounts of ABCB1 mRNA between S- and R-group (P=0.14), although three cases in R-group showed notably high expression levels (Fig. 3D).

Finally, the mRNA expression levels of the DEGs involved in “T cell receptor signaling pathway and PD-L1 expression” and “PD-1 checkpoint pathway in cancer” were compared between S- and R-groups. The expressions of CD3E (P=0.039), ITK (P=0.023), and LAT (P=0.023) genes was significantly increased in R-group compared to S-group (Fig. 3E–G). On the other hand, there were no significant difference in the expression levels of CD28 and ZAP70 genes between the groups (P=0.12 and P=0.14, respectively) (Fig. 3H and I). The expressions of PRKCQ, PD-1, CD8A, and CTLA4 genes could not be compared between the groups by RT-qPCR due to their low expression levels.

DISCUSSION

In the present study, we identified four genes associated with early relapse during first remission induction CHOP-based chemotherapy in dogs with mhBCL.

In the comparison of comprehensive gene expression profiles between S- and R-groups, a total of 179 genes were extracted as DEGs in this current study, of which 169 were increased while 10 were decreased in R-group compared with the S-group. According to the GO term, genes with aberrant expressions in the R-groups were mainly associated with immune response such as chemotaxis, chemokine-mediated signaling pathway, T-cell co-stimulation, and T-cell receptor signaling pathway. Moreover, KEGG pathway analysis showed that extracted 179 DEGs included the genes related to T cell receptor signaling pathway and PD-L1 expression and PD-1 checkpoint pathway.

Previous studies in human DLBCL have demonstrated that the existence of T cells in DLBCL tissues may be identified through the expression of T-cell receptor genes (TCR-β, CD3E), genes downstream of T-cell receptor signaling (Fyn, LAT, PRKCQ), and those specific to cytotoxic T-cells (CD8A, GZMB, PRF, ki67, OX40) [2, 3, 30]. Another report showed the costimulatory interactions between malignant B-cells and T-cells in human DLBCL, and the coinhibitory signals mediated by immune checkpoint seemed to be the main driving force for T cell exhaustion [46]. In addition, it was reported that immune cells in the tumor microenvironment were associated with the outcome in human DLBCL [3, 26, 45]. The studies indicated that patient prognosis was correlated with the abundance of tumor-infiltrating T-cells and their subsets. Furthermore, patients with a high percentage of immune checkpoint-positive T cells had a worse prognosis [3, 26, 45]. Transcriptomic analysis revealed the overexpression of immunosuppressive factors and inflammatory chemokines in relapsed/refractory DLBCL cases when compared to chemotherapy-sensitive DLBCL cases [16]. In veterinary medicine, our previous study showed that the expression levels of the genes associated with immune responses were changed in canine mhBCL when cases acquired resistance to chemotherapy [35]. These previous findings suggest that interactions between tumor cells and the tumor microenvironment, including T cells, may be involved in the response to chemotherapy. The results of the present study also suggested that the early relapse during the first remission induction chemotherapy, which means early acquisition of chemotherapy resistance, in canine mhBCL might be also associated with the differences in the interactions between tumor cells and immune cells such as T-cells.

RT-qPCR was performed to validate the results of RNA-seq using 14 cases in S-group and 9 cases in R-group, of which 16 cases were also used for RNA-seq. As genes associated with GO terms “CCR chemokine receptor binding”, “chemokine activity”, and “chemotaxis”, we examined mRNA expression levels of CCL3, CCL4, and CCL5 genes, and CCL4 showed significantly lower expression levels in R-group compared with S-group. CCL4 gene encodes chemokine (C-C motif) ligands 4, which is one of the chemokines of the CC subfamily and is secreted by a number of hematopoietic cells, particularly those involved in immune responses; macrophages, dendritic cells, and B and T lymphocytes [14]. Normal and malignant B-cells secrete the chemokines CCL3 and CCL4 to attract immune cell such as T cells, for interactions in the tissue microenvironment [6, 7, 9, 24].

We also confirmed that CD3E, ITK, and LAT genes were significantly upregulated in R-group. CD3E gene encodes CD3-epsilon, which is subunit of T cell receptor complex [1, 13]. This complex plays an important role in T-cell receptor signaling and T cell differentiation [8, 24]. In human bladder cancer, patients with lower expression of CD3E in tumor microenvironment had prolonged survival [25]. They also reported that CD3E expression levels are positively correlated with the infiltration of CD8⁺ T cells, CD4⁺ T cells, regulatory T cells, and macrophages in tumor tissues [25]. ITK gene encodes IL2-inducible T-cell kinase, a member of the Tec family of non-receptor tyrosine kinase, acting as an essential mediator of intracellular signal transduction in both T-cells and natural killer cells [15, 27]. ITK plays an important role in the inflammatory processes [49]. While a previous study showed that the mutations in the ITK gene was involved in the development of T-cell lymphomas in human [23], there has been no study that revealed the roles of ITK in B-cell lymphoma. LAT gene encodes linker for activation of T-cells, playing a key role in activation of T-cell signaling pathways following activation of the TCR signal transduction pathway [36]. In human DLBCL, high expression levels of LAT were associated with prolonged prognosis [37]. On the other hand, these genes were up-regulated in R-group in this study, which appears to be contradictory to the results in previous studies.

The findings of our present RNA-seq study indicated elevated expression levels of PD-1 and CTLA4, known as immune-suppressive genes [31], in the R-groups. However, their comparison between groups by RT-qPCR could not be conducted due to their low expression levels. Moreover, it was previously shown that the genes related to inflammatory responses were down-regulated in dogs with chemotherapy-resistant lymphoma [35, 38]. These findings suggest a potential association between decreased anti-tumor immune response and early relapse as T-cells in the tumor microenvironment may experience cellular exhaustion in the R-group. However, we could not clarify the mechanisms that might induce such cellular exhaustion. We need to perform further examinations to investigate the detailed functional differences in the interactions between tumor cells and surrounding immune cells in canine mhBCL tissues.

In the present study, the expression of CCL4 gene was decreased in R-group compared to S-group in RT-qPCR. CCL4, a subfamily of chemokine ligand, is produced by a variety of immune cells including monocytes, B cells, T cells, and dendric cells, and it has been reported that CCL4 play an important role in immune cell migrations and activations [10, 29]. A previous report showed a positive correlation between the expression levels of CCL4 gene and the expression levels of PD-L1 and CTLA4 genes in human renal cell carcinoma [48]. Another previous report showed a negative correlation between the expression level of CCL4 gene and prognosis of human DLBCL patients [44]. These findings were contradictory to those in the present study. The causes of the differences in the findings were unclear, and further studies are needed to elucidate how CCL4 gene expression affects the prognosis of mhBCL cases.

There was no significant difference in ABCB1 expression levels between S- and R-groups in RT-qPCR. However, three cases of R-group showed markedly elevated ABCB1 expression levels, which is similar results to our previous study that compared its expression levels between chemotherapy-sensitive and -resistant dogs with mhBCL [39]. Although it is unclear how significantly the elevated ABCB1 gene expression was associated with the early relapse in these cases, it is possible that the molecular mechanisms of drug resistance might be various among cases. It should be also noted that early relapse in this study is a different phenomenon from the chemoresistance in the previous study.

In the present study, there were differences in the proportions of the cases with dose reductions of any drugs due to adverse events between S- and R- groups. These results indicated that the absent of the dose reductions of the drugs could be associated with the early relapse during the first CHOP-protocol. The previous studies also showed that chemotherapy-induced neutropenia is associated with the prolonged remission duration and survival time in canine lymphoma [42, 43]. However, the number of the cases without dose reductions was small in the cohort included in the present study, and further study is needed using larger number of cases. Furthermore, it is still important finding in the present study that the gene expression profiles were significantly different between S- and R-groups before chemotherapy.

As a limitation of the present study, the tumor samples collected by FNA could contain both tumor cells and stromal cells in tumor tissues, and we could not investigate the precise proportions of tumor cells and other cells in the FNA samples. However, it should be also noted that CD3E, ITK, and LAT genes were possibly expressed by T-cells infiltrated into tumor tissues. A previous study has demonstrated that tumor tissue in canine DLBCL cases typically contains approximately 5% T cells [28]. These findings strongly suggest that the gene expression patterns observed in the RNA-seq results primarily originate from neoplastic B-cells but also reflect the presence of T-cells within the tumor microenvironment. This discussion is further supported by the results of previous studies that conducted RNA-seq using samples obtained from neoplastic lymph nodes [2, 30]. These studies indicated that cases can be categorized based on the expression patterns of T-cell lineage-specific genes, indicating that the samples encompass both neoplastic B-cells and T-cells within the microenvironment. Further studies are essential to investigate the localization of the expressions of mRNAs or protein coded by the extracted genes in tumor tissues by single cell RNA sequencing or immunohistochemistry. Another limitation was that histopathological examinations of LNs could not be conducted and the information on the histopathological classifications could not be obtained in any of the dogs in the present study. Finally, the number of cases in each cohort was relatively small in the present study. Further studies using larger number of samples are needed to confirm the results of the present study.

The results of the present study indicated that the early relapse during the first remission induction chemotherapy, which means early acquisition of chemotherapy resistance, in canine mhBCL might be associated with the differences in the interactions between tumor cells and immune cells such as T-cells. We need to perform further examinations to investigate the detailed functional differences in these interactions within tumor tissues and its associations with early relapse in canine mhBCL.

CONFLICT OF INTEREST

None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of the paper.