Your example $H = \{-1, 0, 1\}$ is not a subgroup of $\mathbb{Z}$ as 1 $\in H$ but 1 + 1 = $2 \notin H$. Newcomers often overlook this fact. In a group, the binary operation (here given by $+$) must be a map from the direct product of the group with itself to the group itself. But the addition on $\mathbb{Z}$ restricted to your subset $H$ will not result in a map
\begin{align*} +_{H \times H} : H \times H \to H
\end{align*},
but rather just a map $H \times H \to \mathbb{Z}$.
One would often say here that H is not closed under addition in technical terms.
For every $n \in \mathbb{N}, n\mathbb{Z}$ is indeed a subgroup of $\mathbb{Z}$, as for each $a, b \in n\mathbb{Z}$, you have that
\begin{align*} a + b \in n\mathbb{Z} \\ -a \in n\mathbb{Z}\\ \text{and the neutral element} \quad 0 \in n\mathbb{Z}
\end{align*}
I believe this already answers the question, but if you are interested to know why the $n\mathbb{Z}$ are the only subgroups of $\mathbb{Z}$, @user85667 gave a short explanation. If you have a subgroup $H$ of $\mathbb{Z}$, then there is an integer with smallest absolute value $n \in \mathbb{N}$ in $H$. Then $H$ must be equal to $n\mathbb{Z}$. Because if there is an element in $H$ that is not a multiple of $n$, then you could divide this element by $n$ and the remainder would be less than $n$. Since $H$ is a group, the remainder would be in $H$, contradicting the fact that $n$ was the element with smallest absolute value in $H$. A generalization of this proof shows why Euclidian Domains are Principal Ideal Domains, you will probably learn about these terms later once you get to Ring Theory.